JPH05504220A - Method and device for identifying and counting currency - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] Methods and devices for the identification and counting of currency by the reflection of light from indicia printed on it Method for automatically identifying and counting currencies of various denominations using features and related to equipment.

has been used. The most rudimentary in this field of technology is the use of dollar bills of specific denominations. They can only handle certain types of banknotes and reject all other types of banknotes. It is a system that allows The most advanced are able to identify, identify, and automate large numbers of denominations. It is a complex system that can be counted dynamically.

Banknote identification systems typically rely on magnetic detection or uses either optical detection. Magnetic detection is usually 7 elite: Abase's The part of the printed mark on the banknote by using a magnetic sensor Detects the presence or absence of magnetic ink in minutes and processes analog or digital processing Based on using the detected magnetic signal as the basis for banknote recognition I'm there. On the other hand, a more commonly used optical detection technique uses a focused stream of light. Changes in the characteristics of light reflection or transmission that occur when a banknote is illuminated and scanned with a trip Based on detecting and analyzing. The identification of banknotes for this is 8! light emitted taking into account the appropriate error in the difference in reflection between individual banknotes of a given denomination of scientific characteristics. Based on comparisons with pre-stored parameters for various face amounts entered. Ru.

51 automatic banknote identification systems! The main problem in implementing The criteria used to properly construct feature patterns and the Analyze test data to identify banknotes with a predetermined pattern The time it takes to compare and consecutive banknotes! ! ! Mechanical feeding and scanning speed and The goal is to get the best compromise between the two. Processing of test data obtained from banknote scanning Even if a microprocessor is used to obtain the sample and the test data The process of identifying the face value of a banknote by comparing it with stored parameters is useful. A limited amount of time is required. Many of the optical stems currently available are banknotes. to obtain a large number of reflection data samples when scanned with a scanning head and and compares these data with the corresponding stored parameters to determine the face value of the banknote. A complex algorithm is used to identify the amount. The traditional system uses face value , to provide sufficient discrimination between denominations, especially those whose reflective patterns are not significantly distinguishable. A relatively large number of optical samples are required for each scan of the bill. large number of data samples This reduces the scanning speed of incoming banknotes and, more importantly, the corresponding Data processing according to the Nu11 algorithm takes a long time. system is identified If the face value is also counted, the processing time will further increase. child As a result, the speed at which banknotes are sorted and counted is increased by 11. This is real time The process must have finished analyzing the data scanned for the banknote. and a banknote is placed across the scanning head and scanned with a rad into the scanning head. Notes must be previously identified and counted as belonging to a particular denomination. This is because it is necessary that the

The main problem associated with traditional systems is the large number of banknotes required to accurately identify banknotes. To obtain a reflection sample of This means that it is constrained to scan. Similarly, scanning also has some fixed There are certain drawbacks. These drawbacks are that the banknotes cannot be moved longitudinally across the scanning head. Disadvantages of long transport paths required for sending banknotes; means associated with the sensing surface of the scanning head to ensure even, non-overlapping alignment; The disadvantages include the added mechanical complexity involved.

The bottom line is that a system that can accurately identify banknotes is expensive and mechanically bulky. It is generally difficult to identify and count banknotes at high speed and with high accuracy. I can't.

Invention summary The main purpose of the present invention is to identify and count currency notes consisting of multiple denominations. An object of the present invention is to provide an improved method and apparatus.

Another object of the invention is to efficiently transfer banknotes of several denominations at high speed and with high accuracy. To provide an improved method and apparatus of the above-mentioned type, which allows identification and counting of That is.

A related object of the invention is to be compact, economical and uncomplicated in structure and operation. An object of the present invention is to provide an improved bill identification and counting device.

Briefly, according to the present invention, the method is adopted for counting banknotes and identifying their face value. The improved optical detection and correlation technique achieves the objectives listed above. this The technology illuminates and scans the banknote along its narrow dimension, approximately in the central section of the banknote. Based on optically detecting the reflective features of banknotes obtained by There is. When the banknote is optically scanned, the light reflected from the banknote is detected and Represents changes in "black" and "white" included in patterns or marks printed on the surface Used as an analog signal.

A series of such reflected signals occurs when the narrow dimension of the banknote becomes parallel to the direction of transport of the banknote. When the banknote moves across the illuminated strip in the The reflected light at the sample point is sampled and digitally processed under the control of a microprocessor. It can be obtained by the following two methods. Therefore, a predetermined number of reflective sensors are applied over the narrow width of the banknote. sample is obtained. The data samples obtained for each scan of the banknote are Due to changes in the "shading" of printed patterns or indicia present on the surface of banknotes Digital processing is applied, including normalization processing to de-emphasize variations. regular The standardized reflection data exhibits a very unique pattern of characteristics for a given denomination. These are used to accurately classify the characteristic patterns for various face amounts. Provide sufficient distinguishing features between the feature patterns of.

For each denomination of a banknote to be detected by using the method described above, A series of master feature patterns are created using the “original” or “new” banknotes. generated and stored. According to a preferred embodiment, for each detectable face value Four characteristic patterns are then generated and stored in system memory. remember The printed pattern is banknote 1; This corresponds to an optical scan performed on the "front" and "back" sides of the banknote along the direction. good Preferably, the method and apparatus for identifying and counting banknotes of the present invention is applicable to U.S. currency. Seven different denominations, namely 1 ton, 2 ton, 5 ton, 10 ton, 2 It is designed to identify $0, $50, and $100. Therefore, 12 A master set of 8 different feature patterns is followed by a system for correlation purposes. stored in system memory.

According to the correlation technique of the present invention, a banknote is scanned under test and the sampled data is processed. For each comparison, the pattern generated by Compare each characteristic pattern to multiple data samples for the pattern being compared. A correlation number is generated that represents the degree of similarity between corresponding ones of the pulls. face value Denomination identification is performed when the scanned banknote is determined to have the highest correlation number by pattern comparison. based on the indication that it belongs to the face value corresponding to the memorized characteristic pattern I'm there. The face value of the scanned banknote is uncharacterized after comparing the feature patterns. The probability that the bi-1level threshold of correlation that must be ).

Essentially, the present invention provides that the banknote travels in either the “forward” direction or the “reverse j direction”≦two directions. Regardless of whether its "front" or "back" side is scanned, multiple Improved optical inspection to clearly identify which of the face value Provide correlation technology. In particular, the present invention provides a method for determining the total amount of banknotes identified at the end of a scanning operation. to track each identified face value for convenient provision at the time of termination. It is designed to be implemented in a system programmed to

Furthermore, according to the present invention, the novel detection and correlation techniques outlined during the Emperor A banknote detection and counting device is disclosed. This device should be counted accepts banknotes, and the banknotes are placed downstream of the curved path for their narrow dimensions; Stacking where banknotes are detected and counted across the scanning head and collected. It has a curved transfer path for transfer onto the station. The scanning head is an optical It works in conjunction with the coda. This optical encoder is used for banknotes (and thus for printing on banknotes). a coherence of light focused below the scanning head (imprinted indicia or pattern) capture a predetermined number of reflectance data samples as it moves across the strip. It's about to start.

Focuses a coherent strip of light of a predetermined size and spreads the light over the illuminated area. using an array of light emitting diodes to obtain a normalized distribution of the intensity of . these Light emitting diodes are placed at an angle and the head scans the desired strip of light. J bifocals on the wide and narrow dimensions of a banknote placed flat across the scanning plane. irradiated Ru. The photoelectric detector is controlled by an optical encoder to obtain the desired reflection sample. Ru.

The beginning of the sampling was determined by comparing the reflection values obtained at the edge of the banknote with no printing at all. t: Based on the change in reflection value that occurs when the outer edges of the pattern meet. It will be carried out. According to a feature of the invention, at least two irradiated beams of different dimensions are provided. The strip is used for scanning purposes. At first, the stamped pattern on banknotes A narrow strip is used to find the starting point, and this narrow strip is representative Mark the starting point of the printed pattern on the banknote and wrap around the printed pattern. It is designed to distinguish the thin border line surrounding it. printed ! ; the rest of the scan of the narrow dimensions, dominated by the detection of the border lines of the banknote; A substantial beam to collect a predetermined number of samples for scanning; a wide strip of light is used. Generating and memorizing characteristic patterns using [original] banknotes; and scan t: if the banknote belongs to one of several predetermined denominations. The subsequent comparison and correlation procedure for sorting is based on the detection correlation technique described above. There is.

Brief description of the drawing Other objects and advantages of the invention will become apparent from the following detailed description when read in conjunction with the drawings. j: Become.

FIG. 1 shows the conceptual basis of the optical detection correlation method and apparatus according to the system of the present invention. It is a functional block diagram showing FIG. 2 shows a diagram for processing and correlating reflection data according to the optical detection and counting technique of the present invention. FIGS. 3 to 8 are block diagrams showing preferred circuit configurations for optical detection. 1 is a flowchart showing the order of operations necessary to implement the outgoing correlation technique; Figures 9A to 9C are obtained by spectroscopically scanning the narrow dimension of banknotes. Figures 10A to 10E are graphs of typical characteristic patterns developed in this invention. By using the gradual movement technique according to the embodiment of the present invention, the correlation pattern between the two FIG. 11 is a graph showing the effect of the optical detection correlation technology of the present invention. No.; is a perspective view of a banknote identification and counting device that is suitable and embodies this technology; FIG. 12 shows a device used to separate banknotes and send them sequentially into a transport path. FIG. FIG. 13 is a side view of the apparatus of FIG. 41 showing the separation mechanism and transfer path; 4 is a side view of the device of FIG. 11 showing details of the drive mechanism, and FIG. 15 is a side view of the device of FIG. FIG. 16 is a plan view of the banknote identification and counting device shown in FIGS. 11 to 14; FIG. FIG. Fig. 48 is a diagram showing the distribution of light created by the turbidity of the optical scanning head; 2 is a plan view of an optical mask used to create scanning strips of various dimensions; FIG. .

Although the present invention is susceptible to various modifications and variations, the specific An embodiment of the invention is shown by way of example and will be described in detail below. However, this There is no intention to limit the invention to the particular embodiments disclosed, but to the contrary The present invention covers all modifications and variations that come within the spirit and scope of the invention as constituted by the appended claims. It is intended to include all changes and modifications.

! ! ! Referring now to FIG. 1, a functional diagram illustrating an optical detection and correlation system according to the present invention. The block diagram of is shown. This system 10 needs to check and calculate It has a banknote receiving station 12 in which a stack of banknotes is placed. accepted The banknotes are separated by a banknote separator that functions to remove or separate the banknotes one at a time. 14, and then the paper is transported along a predetermined transport path in detail. It is relayed by the bill transport mechanism 16 and traverses the optical scanning head 18 to this optical scanning head. At the pad 18, the face value of the banknote is scanned, verified, and calculated. then , the scanned banknotes are transferred to the banknote stacking station 20J and subjected to such processing. digit banknotes are stacked at this station 20 for subsequent removal.

The optical scanning head 18 is arranged such that the banknote 17 is located on the transport path below the scanning head 18. At this time, a coherent beam of light is applied to the paper so as to illuminate a rectangular light strip 24. It has at least one light source 22 directed downward onto the bill transport path. irradiated The light reflected by the strip 24 is detected by a photoelectric detector 26 placed directly above the strip. put out The analog output of the photoelectric detector 26 is converted to an analog-to-digital converter (ADC). Unit 28 converts it into a digital signal, and the output of this unit is used as a digital input. input to the central processing unit (CPU) 30i.

According to a feature of the invention, the banknote transfer path is such that the narrow dimension "W" of the banknote is The transfer mechanism 16 is configured to move the banknotes in a state parallel to the direction. . Thus, as the banknote 17 moves past the scanning head 18 on the transport path, the A heated light strip 24 effectively illuminates the banknote across the narrow dimension rWJ of the banknote. Scan to. Preferably, the transport path is such that the banknote 17 is configured to be scanned along its narrow dimension approximately at its mid-section; There is.

The scanning head 18 scans the banknote as it moves across the illuminated light strip 24. The light reflected from the banknotes was detected in this way, and an analog representation of the changes in the reflected light was generated. This indication shall be the “black mark” contained in the pattern or mark printed on the surface of the banknote. ” and “white” changes. reflected by scanning the narrow dimension of the banknote The change in light indicates the number of denominations that the system of the present invention is programmed to handle. It is useful as a measurement value for identifying the forehead with high accuracy.

A series of such reflected signals detected may be due to the narrow dimension of the banknote or the selected width of the banknote. The obtained analog signal is converted into digital signal under the control of CPU 30. digitized to generate a predetermined number of digital reflection data samples. Then these Add digitization processing to the data sample. This process involves Processing sample data and determining the shading of the printed pattern on the surface of banknotes. A normalization routine for t: is included to smooth out variations due to changes. in this way The reflection data is digitized and normalized as: exhibiting a very distinctive characteristic pattern and differing denominations as detailed below. Provide sufficient discriminative features between the feature patterns for.

There is no exact distance between reflection samples obtained by scanning the narrow dimension of successive banknotes. To ensure compliance, the initiation of the reflection sampling process is preferably initiated by CPU3. 0 through an optical encoder 32. This optical encoder is suitable for banknote transfer machine 16 to closely track the physical movement of the banknote 17 across the scanning head 18. Ru. More specifically, the optical encoder 32 detects whether the banknote is being relayed along the transport path. related to the rotational movement of the drive motor that produces the movement applied to the banknote. Ru.

Furthermore, especially when a banknote is scanned by the scanning head 18, there is an active gap between the banknote and the transport path. contact is maintained. The optical encoder cannot be driven under these conditions. A light strip generated by a scanning head by monitoring the rotational movement of a dynamic motor l: The movement of banknotes can be tracked in detail.

The output of the photoelectric detector 26 is monitored by the CPU 30, and the presence of a banknote under the scanning head is first detected. Then, the starting point of the printed pattern on the banknote indicated by the thin border line 17A is detected. put out The border line 17A surrounds the printing on the banknote. Once the border line 1 7A, the banknote 17 moves across the scanning head 18 and its narrow dimension The reflection sample obtained from the output of photoelectric detector 26 when scanned along Optical encoders are used to control timing and number.

Detection of border lines constitutes an important step and achieves improved identification efficiency. This is a good This is because the decorative line serves as an absolute reference point for starting sampling. Ru. If you intend to use the edge of the banknote as a reference point, the relative sampling point Misalignment may occur. This requires relatively large tolerances during printing and cutting of banknotes. Because the distance from the edge to the border line varies depending on the banknote due to tolerance tolerance, It's a good thing. As a result, i['! between sample points of consecutive banknotes. Ensure fC response This makes it difficult to maintain data and has a negative impact on identification efficiency.

Provides control of the sampling process for the physical movement of the banknote across the scanning head. Using an optical encoder to detect a predetermined delay prior to the start of the sample It is also advantageous in that it can be used to detect the a<am9 line. encoder The delay is due to the narrow width of the banknote containing the most distinctive printing for different denominations. You can adjust the bill so that it is scanned across only the legal segments. Wear.

For example, in the case of U.S. currency, when scanning across the center section of the narrow dimension of the bill, the paper The approximately s, oscm (approximately 2 inches) section in the center of the bill shows the face value of various U.S. currencies. Based on the correlation technique of the present invention, sufficient data is provided to distinguish between ing. Therefore, optical encoders can be used to control the scanning process, and As a result, the reflection sample is detected for a set period of time and for a predetermined time after detecting the border line. This limits the scanning to the desired central part of the narrow dimension of the banknote. Set.

Optical detection correlation technology detects a “new” instantaneous value for each denomination of the currency to be detected. generate a set of master feature patterns using the ``original'' banknotes. 2 is based on using the process described above. According to a preferred example, four A characteristic pattern is generated, preferably an EEFROM (Electrical Eraser/Writer). 34 (see Figure 1) (see Figure 1). It is stored for each face amount that can be withdrawn. The characteristic pattern of each banknote is determined by optical scanning, i.e. It is carried out on the front and back sides of banknotes, and the pattern printed on the banknotes is A predetermined number of reflection samples taken along both the "forward" and "reverse" directions. generated in response to the process of obtaining the file.

When applying the technology of the present invention to US currency, for example, characteristic patterns may be generated. Seven different denominations of U.S. currency: $1, $2, $5, $10, Stored for $20, $50, and $100. Therefore, the following correlations A master set consisting of 28 different feature patterns was created as a system memo. stored in the memory. Once you memorize the master feature pattern, scan the banknote under test 28 pre-stored master patterns The CPU 30 compares the degree of correlation, that is, the comparison t; For each comparison, the correlation number represents the similarity between the corresponding one of the data samples. Occurs every time.

It can be seen that CPU30 has the highest correlation number obtained by pattern comparison. Check the face value of the scanned banknote as it corresponds to the memorized feature pattern that was scanned. is programmed to do so. Failure to characterize the face value of scanned banknotes Eliminate the possibility and at the same time confirm that the counterfeit banknote belongs to a valid face value In order to reduce the possibility of A two-level threshold of correlation is used as the basis for making calls.

By using the detection correlation method described above, CPU 30 can banknotes belonging to a particular currency denomination as part of a given set of banknotes quantified for Count the numbers and determine the total amount of currency represented by the banknotes scanned during the scanning batch is programmed to do so. Furthermore, the CPU 30 is connected to the output unit 36. This output 36 is counted and displays the number of banknotes and the currency face value. The breakdown of the banknotes and the total face value of the currency represented by the banknotes counted. It is designed to give a meter. The output unit 36 converts the displayed information into the desired form. It is also possible to open the print door using a formula.

Referring now to FIG. 2, there is shown processing of reflection data according to the system of the present invention. A preferred circuit arrangement for correlation is shown in block diagram form. Shown here The CPU 30 includes an optical encoder 32, a photoelectric detector 26, and a memory unit. Various input signals, including signals from unit 38, are accepted and processed. memory uni The bit 38 is a static random access memory (RAM) or erasable programmable memory. memory unit 3, which is preferably an EPROM read-only memory; 8C; stores a correlation program, and creates a pattern based on this program. is generated and the face value of the test currency is confirmed; the test pattern is stored in the memory card. Compared to the Star Program. Crystal 40 is used as the time reference for CPU30. This is useful, in addition to an external reference voltage V□, as explained below. Peak detection of the detected reflection data is performed based on this voltage.

The CPU 30 receives the timer resetting signal from the resetting unit 44 and sends the timer resetting signal to the resetting unit 44. The unit 44 receives the output voltage from the photoelectric detector 26 as shown in the second A[J]. and set the second voltage to a preset voltage threshold, typically 5.0 volts. When the detector 44A compares and detects the radiation value corresponding to the presence of paper, it becomes "high". Provide a go reconfiguration signal. More specifically, reflection sampling If no bill is placed under the illuminated light strip (24 in Figure 1) It is based on the assumption that no light is reflected to the photoelectric detector. In these conditions, light The output of the electric detector represents a "dark" or "zero" level reading. Edge of bill scanned first When placed under the head and under the light strip 24, the output of the photoelectric detector changes to a "white" reading, typically set to have a value of about 5.0 volts. Ru. When this happens, the reconfiguration unit 44 provides an r high J signal to the CPU 30J. Record the start of Run 1 procedure.

According to a feature of the invention, the thickness of the illumination light strip produced by the light source in the scanning head is set relatively small at the first stage of the scan when thin border lines are detected. It is. By using a narrow slit, the detection sensitivity of reflected light is increased. , allowing small variations in the "ash" level reflected from the banknote surface to be detected. Ru. This means that the fine border line of the pattern, i.e. the starting point of the printed pattern on the banknote, is This is important for ensuring accurate detection. After detecting the border line, next For complete scanning over narrow dimensions of banknotes and for quickly obtaining the desired number of samples. Ru! Reflection sampling is then performed based on relatively wide light strips. actual Using a wide slit for sampling smoothes the output characteristics of the photoelectric detector. Realizes a relatively large analog voltage. This is an accurate display of the detected reflection value and processing.

Returning to FIG. 2, the CPU 30 passes the output of the photoelectric detector 26 through the peak detector 46. Process it. The peak detector 46 essentially collects the output voltage of the photoelectric detector. and hold the highest or peak voltage value encountered after the detector is activated. functions. Additionally, the peak detector is adapted to determine a normalized voltage; Based on this, the border line of the pattern on the banknote is detected. The detector 46 is a photoelectric An ADC 48 for digitizing the output of the detector and a preselected output from the power supply 42. A digital signal for reconverting the signal into analog form based on the reference voltage V□, and an analog converter (DAC) 50. The output of DAC5O is code converted 52 to a voltage divider 54, which divides the input voltage to a predetermined peak value. The voltage is lowered to a normalized voltage V, which represents the determined ratio. The voltage V, is the voltage at which the banknotes are printed. Not 1#! , a thin border line due to the "high" reflection value due to scanning the area. When the peak detector represents a change to a relatively low "gray" reflection value of the bush encountered with is based on the percentage drop in the output voltage of the peak detector. Preferably normalized Voltage VS is set to be about 70% to 80% of the peak voltage.

The normalized voltage V, is supplied to a curve detector 56, which includes a photoelectric detector. 26 incoming instantaneous outputs are provided. Curve detector 56 receives two voltages as its inputs. The side compares, signals and generates DffiT. This signal normally remains “low”. and towards "high" when the edge of the bill is scanned. Signal L oar is photoelectric detection Predetermining the peak voltage of the photoelectric detector to the point where the input and output of the detector is represented by the voltage V, When the percentage reached is reached, it goes to "low". Thus, a signal is sent and the machine is directed to "low". This is an indication that the border line of the banknote pattern has been detected. Ru. At this point, the CPU 30 is under the control of the encoder 32 (see section 2) Begin reflection sampling and move the bill across the illuminated light strip until its width When scanned along the central section of narrow dimension, the desired predetermined number of samples is obtained. It will be done.

By scanning a "new" banknote when a master feature pattern is generated The resulting reflection samples are entered into corresponding designated sections within system memory 60.

The system memory is preferably EEFROM. Sample input is required If so, it is done through the buffer address latch 58. Preferably a master putter This is caused by scanning the "new" banknote multiple times, typically and the corresponding deleno before memorizing the average value as representing the master pattern. – Obtain the average value of the samples. Reflections obtained by scanning a test banknote during banknote identification EEFROM under the control of a correlation program stored in memory unit 38. 60 and, again, an address latch. Compare sequentially through 58.

Referring now to FIGS. 3 to 7, these figures show the optical detection system of the present invention described above. A flowchart illustrating the sequence of operations involved in implementing the correlation technique is illustrated. Especially the first Figure 3 shows the presence of a banknote under the scanning head and the presence of a border line on the banknote. Indicates the order required to detect and. System program shown as "trigger" This division begins at step 70. In step 71, the banknote start interrupt is executed. A decision is made as to whether or not to proceed. This means that the system will not be able to detect the presence of banknotes. The state is ready and set, i.e. it defines what is happening.

If the answer in step 71 is yes, go to step 72 and continue with this step. In the step, the operation is performed based on the resetting procedure described above with respect to the resetting unit 44 of FIG. Make sure there is a banknote under the scanning head.

If the answer in step 72 is yes, i.e., a banknote is present, then step Moving to step 73, in this step the peak value is increased to a predetermined percentage thereof. A test was conducted to find out whether border lines were detected based on the reduced It will be done. As mentioned above, the decrease in the peak value causes the signals LI, , A to go "low". It is shown by. If the answer to step 73 is no, then the program continues the loop until a border line is detected. If the answer to step 72 is no If the banknote does not exist, the banknote start interrupt is re-established in step 74. and the program returns from the interrupt at step 75.

When it is determined that a border line has been detected in step 73, the A/D completion interrupt is used. Enable step 76 is called, whereby the analog-to-digital conversion This indicates that the conversion can be subsequently performed at desired time intervals. then , the output of the optical encoder at step 77 at which the first reflection sample is to be obtained. Decided in relation to. In step 78, the rSTARTA By recalling the routine denoted as 2DJ, the detected reflected sun Capture and digitization of pulls will be undertaken. Banknotes upon completion of the digitization process A termination interrupt is enabled in step 79, and this step Reconfigure the system to detect the presence of subsequent banknotes. Then step 80 The program returns from the interrupt. Bill start interrupt not performed in step 71 If so, it is determined in step 81 to find out whether a bill end interrupt has occurred. will be done. If the answer at step 81 is no, the program Step 85 returns from the interrupt. If the answer is yes in step 81, then Step 83 is called, in which the bill start interrupt is activated and the existence of the bill is The resetting unit that monitors the presence of the banknote is ready to determine the presence of the banknote in step 84. It will be reconfigured as shown below. The program then returns from the interrupt in step 85. Ru.

Referring now to Figures 4A and 4B, these figures include the 5TARTA2D route. The routine for starting the process and its digitization routine are shown respectively. . In FIG. 4A, the 5TARTA2D routine is started at step 90. sample point which gives an indication of the sample obtained and digitized at a given time. Initialize the interface. Then, in step 91, an analog-to-digital conversion is performed. Make specific channels interruptible. Allow the first sample to be digitized. The interrupts that are enabled are enabled in step 92, and the main program Called again at 93.

FIG. 4B shows the sequential steps required for the analog-to-digital conversion routine 70- This is a chart and is designated as rA2DJ. This routine is step 100 Start with. The samples are then adjusted to maintain an indication of the number of remaining samples to be obtained. The pull pointer is decremented in step 101. In step 102, the current sample Read the digital data corresponding to the output of the photoelectric detector. this data is converted to its final form in step 103 and divided into predetermined memory segments. It is stored as XIN in the memory card.

Next, in step 105, it is determined whether a desired predetermined number of samples rNJ have been obtained. Do a second check. If the answer is no, step 106 is called. In this case, an interrupt is enabled to allow digitization of subsequent samples. The program steps from the interrupt to complete the remainder of the digital program. Return at 107. However, if the answer in step 105 is yes, then If the desired number of samples have already been obtained, a flag indicating this will appear in the step. is set at 108 and the program returns from the interrupt at step 109.

Referring now to Figure 5, this diagram shows the mathematical steps required for the correlation process. The sequential steps necessary to carry out the routine shown as rEXECJ are shown. It is. The routine begins at step 110. In step 111, All interrupts are disabled, and the CPU is initialized at this time. Step In step 112, constants associated with the sample process are set, and in step 113, constants associated with the sample process are set. communication protocols and related information, if any, for exchanging processed data. As a result, the number of ports, interrupt mask, etc. are defined.

In step 114, the resetting unit indicating the presence of banknotes is to be scanned. Reset to detect the presence of the first banknote. At step 115, the banknote starting discount The interrupt is enabled, leaving the system waiting for the first incoming bill. then , all other associated interrupts are enabled in step 116. This is two At this point, the initialization process is complete and the system is ready to start scanning banknotes. This is because As to whether all the desired number of samples were actually obtained. A check is then made in step 117. If the answer to step 117 is no If so, the program loops until it gets a yes answer. Jesus' answer When the value is obtained, step 118 is called, and in this step 7 lag is set. and indicates the start of the correlation procedure.

According to the present invention, the digitized reflection values are processed and pre-stored in the same format. Quantitative correlation procedures are used to form a form that is conveniently and accurately compared with corresponding values. used. More specifically, as a second step, the The average value X of a set of digitized reflection samples (comparing n samples) is as follows: Get it as below.

n　Xl X-Σ □ ・ ・ ・ (1) 1 “On” Then set the normalization factor “σ” as normalized by the total number of samples. , is determined to be equal to the sum of squares of the differences between each sample and the mean value. More specifically, σ −Σ□・・・(2) i ~On The final step is to calculate the sample and above as defined in the equation below. Obtain the difference between the average value and divide this by the square root of the normalized 7 actangma "σ" Normalize each reflection sample by

Using the above correlation equation, following the normalization process, the test pattern and master – Pairing with the test pattern and any master pattern that exists between the If the total number of products in corresponding samples is equal to the total number of samples, then the pattern is the same. If so, a correlation equal to l can be obtained. Otherwise, from 1 A small value can also be obtained. Therefore, the correlation number, i.e. the normalized sample in the test pattern. factor obtained by comparing the sample with the memorized master pattern clearly displays the degree of similarity or correlation between these two patterns.

Preferred 5 of the present invention! According to the example, the banknotes are digitized and normalized for scanning. The predetermined number of reflected samples is selected to be 64. This is in binary Using samples of even higher orders of magnitude (128, 256, etc.), the correlation procedure described above yields a correspondingly increased identification efficiency for the greater processing time required to implement I know from experience that it doesn't. Furthermore, smaller than 64 such as 32 Using large binary digits significantly reduces identification efficiency.

To facilitate correlation, correlation factors can be conveniently expressed in binary terms. example For example, in the preferred embodiment, 1 of 7 actors when 100% correlation exists is binary. It is expressed in modulo 210, which is equal to 1024 in decimal. closest to 1024 Identify the test patterns that most closely correspond by identifying the comparisons that yield the correlation numbers. within the test pattern using the procedure described above to determine the memorized pattern of The normalized samples of 28 @ master features stored in system memory Compare each of the patterns.

A feature of the invention is that a two-level threshold of correlation is met before making a particular call. I need to add it. More specifically, the correlation procedure combines the test pattern with a stored pattern. equate the two highest correlation numbers obtained by comparing one of the turns It is designed to be set. At this point, we can define the minimum threshold of correlation with these two correlation numbers. need to be satisfied. It turns out that a correlation number of about 800 serves as a good threshold. We know from experience that above this correlation number, a yes call will be made with a high degree of certainty. , below this correlation number, if it corresponds to any pattern among the stored patterns, The display of the test pattern is a failure. As the threshold level of Wc2, the coal shear is defined as the minimum separation between the two highest correlation numbers. This allows the test pattern to pattern corresponds to one or more stored master patterns within a predetermined correlation range. A yes call is made with certainty only when there is no such thing. Preferably, the first between the correlation numbers The separation is set to be between 100 and 150.

Referring now to FIG. 5, the correlation procedure begins at step 119, which step , the routine labeled rPROcEssJ is called. this routine The steps necessary to carry out the procedure are shown in FIG. 6A. @6 Ae; lJ is step 1 The routine starts at 30. In step 131, the average value X is converted into equation (1). Calculate based on In step 132, the sum of the squares is calculated according to equation (2). It will be done.

In step 133, the reflection samples are expressed in integer form for further processing. Convert the digitized value of to floating point format. In step 134, all A check is made as to whether the sample has been processed and if the answer is yes The routine ends at step 135 and calls the main program again. Ste If the answer at step 134 is no, the routine returns to step 132. The above calculation is repeated in step .

At the end of the PROCESS routine, the program returns to routine EX at step 120. Return to the EC and confirm here that all digitized reflection samples have been processed. The flag indicating this is reset. Then, in step 121, r! Displayed as MGCALJ The specified routine is called. Step 6 describes the steps necessary to carry out this routine. Figure 3 shows this. FIG. 6B shows the routine beginning at step 140. Step Step 141 calculates the square root of the sum of squares, as computed in routine PROCESS. is calculated according to equation (2). In step 142, the routine PROCESS Using the floating point value calculated in step 141, the equation (3 ) - normalize accordingly. In step 143, all digital samples are processed. Check whether it was done. If the answer to step 143 is no, then , the program returns to step 142 and continues the conversion until all samples have been processed. Let's go.

At point 2, the answer in step 143 is yes, and the routine returns to the main program. The process returns to step 144.

Returning to the 70-chart of FIG. 5, the next step to be performed is step 12. 2, and the routine labeled rCORRE LJ is called at this step. It will be done. The steps necessary to carry out this routine are shown in FIG. Figure 7 shows this le. Indicates that the routine starts at 150. In step 151, the correlation result is In step 152, the test pattern is is compared with the first one of the patterns. In step 153, up to this point The first call corresponding to the highest correlation number obtained is determined. Step 15 4, the second call corresponding to the highest correlation number of $2 obtained up to this point is determined. determined. In step 155, the test pattern is compared with all master patterns. A check is made to see if this has been done. If the answer is no, then Chin returns to step 152I and repeats the comparison procedure in this step. all ma When the star program is compared to the test pattern, the Ste knob 155 will turn YES. The routine returns to the main program number; step 156.

Returning again to FIG. 5, at step 123 a 7 lag signal indicates that the correlation procedure is complete. and step 124 is called, in which rsEROUTJ and The displayed routine is started. Steps 118 and 123 add up to the entire correlation procedure. The setting of flag TP2, which has the main function of measuring the necessary processing time, and Please note that this is about settings. These steps monitor processing time You can omit it if you don't. Steps required to perform routine 5EROUT is shown in the eighth factor. The wls diagram shows that this routine begins at step 160. show. In step 161, the face value corresponding to the $1 call is changed to ASCII format. will be converted and displayed. In step 162, the correlation number corresponding to the kJc1 call is Convert to ASCII format and display.

In step 163, the face amount corresponding to the second call is converted into ASCII format, indicate. In step 164, the correlation number corresponding to the second call is converted into ASCII format. Convert and display. The routine then returns to the main program. In this regard, mail In-program, the correlation step is completed and any related calculations that identify the face amount are A number is made with the relevant results and displayed along with the corresponding call and correlation numbers. child After completing the correlation-display procedure, the system starts the scanning process for the next incoming note. Ready for.

In implementing the optical detection correlation technique of the present invention, (i) the sampling and correlation process; (ii) to control the overall functioning of the entire system; It is noted that a microprocessor unit may be provided. like this In such embodiments, the entire processor unit preferably has an identified face value. and any related counting results. In this method, Ruchi 5EROUT (see Figure 8) is used for transmitting the amount on the M side of banknotes and for sampling. - from the correlation processor unit the entire processor unit for display; It only requires a call for information to the client.

next,! Referring to Figures 9A through 9C, these figures show a dollar bill as Surface 1; test pattern produced when scanning forward along a two dollar bill; The test pattern produced when scanning in the opposite direction on a surface, and the 100 dollar bill. Each test pattern is shown as it occurs when scanning forward around the surface.

In 9th A[El to 9th CI1m, samples are taken to clarify the test pattern. Unlike the preferred method, which uses only 64 + i Figures 9Ae to 9C are created by using sample 2. I would like to draw attention to this. For these three test patterns, the corresponding samples The recorded differences that exist between the files are indicative of a high degree of accuracy, and thus The face value is called using the optical detection correlation procedure described above.

The optical detection correlation technology described above allows for highly accurate identification of pre-programmed The optical detection correlation technique described above digitizes the sampled reflection values and based on the relatively small processing time to compare the values of Ru. This method scans a banknote, normalizes the scanned data, and has the most distinctive printed mark: the compared sample points of the banknote section used to generate master patterns with direct correspondence between them. Iku A relatively small number of reflective samples to be able to properly distinguish between a few face values is required.

The main advantage of this method is that the banknote does not have to be scanned along its wide dimensions. That is to say, yes. Traditional systems typically do not accurately identify face value11. How to scan wider dimensions to obtain the larger number of samples needed to has been forcibly adopted. Furthermore, by reducing the number of samples, Additional comparisons can be made during the time available to scan the amount of banknotes that have been scanned. Shorten processing time to a certain extent. More specifically, as mentioned above, the test pattern The comparison can now be made with at least four memorized master feature patterns. Therefore, the system can move the banknote in the "forward" direction or along the "front" or "back" side of the banknote. It is made to be able to identify banknotes that are scanned in the "reverse" direction.

Another benefit of the reduced processing time achieved with the detection correlation method of the present invention is that i.e. identified as not corresponding to any of the stored master feature patterns. either stopping the transfer of banknotes or diverting such banknotes to another stacker bin. This means that the reaction time required to do so is correspondingly shortened.

Therefore, this system ensures that the scanned pattern does not correspond to any of the master patterns. It can be conveniently programmed to set a flag if it does not. child Identification of conditions such as can be used to stop the banknote transport drive motor for the mechanism. Ru. Since the optical encoder is associated with the rotational movement of the drive motor, the state before stopping synchronization can be maintained between the current state and the state after the stop. Implementation using two processes described above In the example, information about the identification of "counterfeit" banknotes is transmitted to the entire processor unit This in turn controls the drive motor appropriately.

The accuracy of the correlation procedure and the face value identified by this procedure depends on the test pattern. The correspondence between the reflection samples of and the corresponding samples of the stored master pattern directly related to the degree of Thus, the shrinkage of "used" banknotes is due to their narrow width. The dimensions should be adjusted accordingly to correspond to such spent banknotes of a given face value. This may reduce the degree of correlation with the master pattern. Ganari user For older banknotes, this reduction in dimensions occurs in both the narrow and wide dimensions of the banknote. happen. The detection correlation technique of the present invention is relatively unique from changes in the wide dimensions of banknotes. Remaining upright and decreasing along the narrow dimension, the "shrunken" bill will lie sideways to the scan. By realizing the relative displacement of the reflection sample obtained when it is cut and transported, may affect the correlation factor.

In order to absorb the effect of such shrinkage in narrow dimensions, that is, to make it zero, the master pattern is Divide the test pattern that does not correspond to any of the turns into predetermined sections and In order to identify the face amount, the samples in successive sections are moved progressively and stored. By using the progressive movement method, which is again compared to Tabata, the above-mentioned correlation technique It is better to change the technique. This gradual movement causes the bill to shrink along the narrow dimension. It has been empirically determined that this is an effective countermeasure against sample deviation caused by There is.

The effect of gradual movement is maximized by the correlation patterns shown in Figures 1aArM to 100. is also often exemplified. For clarity purposes, the example pattern is shown for each bill scan. Therefore, using 12g samples compared to the preferred use of 64 samples. I'm exhausted. The xoArM is the master corresponding to the test pattern (shown in bold). - shows the correlation between the patterns (shown by thin lines). degree of correlation between two patterns It is clear from Figure 10A that the correlation factor is relatively low, exhibiting a correlation factor of 606. Ru.

Here's how to increase the correlation between these patterns by using progressive movement. , the correlation is considered at the reference points denoted as A to E along the axis that constitutes the number of samples. This is best exemplified by considering Correlation by “-times” gradual movement The effect brought about is shown in FIG. 10B. This diagram shows the test pattern in Figure 10A. It shows "-times" movement of . This creates two equal samples with 64 samples each. This is done by dividing the test pattern into segments. The first segment is The second segment is held without any movement, whereas the second segment consists of 7 apertures of l data samples. Only Kuta has moved. Under these conditions, the correlation factor is The reference point located, specifically point E, is improved.

Figure 10C shows the effect obtained by performing "triple" incremental movements. test The segmentation of the pattern is thereby moved in three stages. This is butter, This is done by dividing the entire area into three approximately equally sized sections. Category 1 is not moved, and section 2 is moved by one data sample (as shown in Figure 10B). and Segment 3 has been moved by 7 actors of 2 data samples. Move “three times” It can be seen that the correlation factor at point E increases further.

Based on the same criteria, Figure 10D brings the correlation by "three" progressive movements. In this case, the entire pattern is first divided into four approximately equally sized Divide into sections. Then, keep section 1 without moving and take one data sample of section 2. Move section 3 by 2 data samples, move section 4 by 3 data samples. Move only the file. In this state, the number of correlated 7 actors at point E is further increased. I understand.

The 10th EUgJ shows the effect that a “four” move has on the correlation; In this case, divide the pattern into five approximately equally sized sections. ＃L's first four wards The minutes are moved according to the method of performing the "three" movement in Figure 10D, but the third The minute is moved by 7 actors of 4 data samples. From Figure 10E, the phase at point E This means that the relationship is increased until it approximately matches the superposition of the compared data samples. It is clear that

The advantage of using the gradual movement method is that only a set amount of data samples Quite different from moving, the results obtained in the first segment of the pattern as a result of moving that the improvement in the correlation that occurs is not offset by subsequent shifts in the test pattern. It is. This is the degree of correlation for sample points that fall within progressively moved partitions. It is clear from the above figure that the amount increases accordingly.

More importantly, the gradual shift reduces the total correlation factor obtained by pattern comparison. Achieving a significant increase in data. For example, the correlation 7 actor (*10 (see Figure A) is increased to 681 by moving "- times" as shown in Figure 10B. . The "three" moves shown in Figure 10C increase the number of correlations to 793, resulting in $10D Figure I The "three" moves shown here increase the correlation number to 906, and finally the 10th Et! Shown in l. The "four" moves increase the total number of correlations to 960. Using the method described above, The narrow dimension has shrunk considerably, and the correct amount when correlation is performed without any movement. using an incremental transfer method to remove spent banknotes that cannot be accurately identified as belonging to a denomination; Preferably, ``three'' or ``four'' movements are employed, which allows for a higher degree of accuracy. Can be identified.

Referring now to FIG. 11, this figure shows a currency identification system embodying the principles of the present invention. f+1 and a device R21o for counting are shown. This outfit flI, left @W 1214, right side! ! 216, later! 2218, with reference numeral 220 throughout. It has a housing 212 including a top surface. The device consists of a generally vertical front section 22 4 and a forward sloping section 225, the front sloping section 225 having a , various control switches for operating the device and associated display means are installed. 226A and 226B.

Accepts a stack of banknotes 22B that must be graded according to their face value The input bin 227 is placed on the top surface 220 by the downwardly inclined support surface 229. It is formed. A pair of vertically disposed side walls 230 and 2 are disposed on the support surface 229. 32 are provided and these gIJ walls are separated from each other by a vertically disposed front wall 234. connected. ! ! 230.232.234 cooperates with the inclined surface 229 to A stack of 228 constitutes a positioned enclosure.

From the input bin, the banknotes travel along a transport path with three sections. The route of travel is , an input path in which the banknote moves along the first direction in a substantially flat position; a curved guideway for accepting and guiding the direction of movement to a second different direction; and the banknote is placed downstream of the curved guideway along a second different direction in a flat position. having an output path for moving across a banknote identification means, as described in more detail below, located therein. . In accordance with the improved optical detection correlation technique of the present invention, the transport path is is transferred along the input path, the curved guide path, and the output path, and the narrow dimension r of the banknote is W'' are stacked so that they are always maintained parallel to the transfer path and direction of movement. This consists of:

The forward inclined section 225 of the document handling device 210 is located on the side! 214.216 The platform surface 235 has a banknote identification plate. Banknotes that have been processed by other means are consolidated. Processed banknotes are stacked for removal. The stacker plate 242 is designed to receive and deliver the paper to the stacker plate 242 . more specific In this case, the plate 235 has an associated angular surface 236 and an opening 237.23. 7A, and from these openings the corresponding pairs of stacker wheels 23"8 , 240 extend outwardly, respectively. these The stacker wheel of is arranged around the angular surface 236 and the side walls 214 and 21 Rotating around the stacker shaft 241 suspended across 6. is supported by The flexible blades 238A, 240A of the stacker wheel are In cooperation with stacker platform 235 and opening 237. Take out the banknotes. These blades then move these bills onto the stacker plate 24. It operates in such a way that it sends out two fights and two. A stacker plate 242 is coupled to the angular surface 236; This stacker plate is also provided with stacker wheel openings, and the wheels fit into these openings. stands out from During operation, the banknotes delivered to the stacker platform 235 are It is picked up by the visibility blade, passes between a pair of adjacent blades, and connects these blades. The rads combine to form a curved enclosure into which the supports the banknotes as the stacker wheel rotates and supports the stacker platter. Serves as a means for transferring from platform 235 onto stacker plate 242. vinegar Mechanical configuration of tacker wheels and flexible blades on stacker wheels , and how these stacker platforms and plates work together. and therefore will not be described in detail herein.

The banknote handling and counting device 210 is configured to take out banknotes, that is, to load them into the input bin 227. A means is provided for "peeling off banknotes one by one from a stack of banknotes." Ru. To provide this pulling and stripping action, a feed roller 246 is attached to the drive shaft 24. The drive shaft 247 is rotatably mounted around the side wall 214 . It is supported between 216 and 216. The supply roller 246 is tilted below the input bin 227. The input bin protrudes through a slot provided in the surface 229, and the input bin constitutes an input path. to be accomplished. The supply roller 246 has a relatively high friction support on at least a portion of its periphery. It is in the form of an eccentric roller with surface 246A. The surface 246A is the rotation of the roller 246. It is adapted to engage the lowest banknote in the stack of banknotes 228 when rolling. t Therefore, the lowest point along the supply direction indicated by arrow 247B (see FIG. 13) The movement of banknotes begins. The eccentric surface of the supply roller 246 essentially Rotate the stack of banknotes once per revolution, moving and loosening the lowest banknote in the stack. “Shake and push.” This results in the advancement of the bottom bank note along the feeding direction. is facilitated.

A cab supported to rotate around a capstan drive shaft 249 The action of the feed roller 246 is supplemented by the provision of a drum or drum 248. Ru. The capstan drive shaft 7) 249 is supported between the side walls 214 and 216. ing. Preferably, capstan 248 has a smooth surface and is made of rubber or BJJLJ Centrally located molded from a material that provides friction, such as plastic It has a friction roller 248A. This friction roller consists of a pair of capstan rows. It is sandwiched between rollers 248B and 248C, and the outside of these capstan rollers is At least a portion of the circumference is provided with a surface 248D that provides high friction.

Friction surface 248D is similar to friction surface 246A provided on the supply roller; This causes the capstan roller to frictionally move the lowest note along the feeding direction. can be done. Preferably, the rotational movement of capstan 248 and supply roller 246 is The friction surface provided around the carb tongue and supply roller rotate together. is synchronized so that the complementary induces frictional contact.

The capstan 248 and the supply roller 246 are rockingly driven and forward. Two input bins to ensure effective contact with the banknotes in the process of being advanced. supported on ejection arms 254A, 254B, these ejection arms The support shaft 256 is supported between the side walls of the support shaft 256. held. The take-out roller is designed to rotate freely around the take-out arm. If there is no bill in contact with the capstan 248, the friction roller 248, thus inducing rotation in the opposite direction to this friction roller.

However, if a bill is present and in contact with capstan 248, The delivery roller comes into contact with the leading edge of the banknote, and the rotational movement of the roller is inhibited. exerts a downward force on the banknote. As a result, capstan roller 248B .

Advancing caused by contact between friction-providing surface 248D on 248C The action is enhanced, which allows the banknotes to be removed one at a time from the banknote stack 28. Peeling off is made easy.

Between the take-out arms 254A and 254B, the support/yaft 256 is connected to the separation arm 26. 0, this separation arm has a stationary strip at its end remote from the shaft. The stripping roller supports the stripping roller 258, and the stripping roller is attached to the stripping roller. A friction surface is provided to provide frictional resistance on the banknote. The separation arm is The arm is attached to move in an arc around the foot 256, and this arm is is spring loaded to abut downwardly on the capstan with an amount of force.

In operation, the take-out rollers, due to their free-rotating nature, With the rotational movement of the friction roller 248A until it encounters the leading edge of more than one bill, Rotate. When it encounters a banknote, the rotational movement of the take-out roller stops and the banknote is removed. The leading edge is forced into positive contact with a friction-providing surface on the circumference of the capstan roller. be made to touch. This effect separates the bottom banknote from the capstan from the rest of the banknotes. This is to forcibly separate them along the direction of rotation. At the same time, separation shoe 2 58 is also lowered onto any of the bills propelled forward by the capstan roller. to come into contact with someone.

The tension acting on the take-out arm 254A is not exerted on such propelled banknotes. The downward force applied is selected to allow only one bill to move forward. Ru. Due to the contact created between the take-out roller and the capstan roller, If one or more banknotes are ejected, the spring-loaded shoe The downward force exerted thereby prevents these bills from moving further forward. must be sufficient to The spring-loaded tension on the pick-up arm To supplement the banknote peeling action created by the output roller and capstan roller. Adjustments are made to the kari to control the downward supporting force exerted by the nyu. I can do two things. Thus, the rotational movement of the capstan allows two or more banknotes to be At the same time, the possibility of being propelled forward is greatly reduced.

The bill transfer path rotates forward along an input path formed by the front section of the ramp 229. a capstan for accepting banknotes propelled into frictional contact with a capstan It has a curved guide path 270 provided in front of the tongue 24B. The guide route 270 is The curved section 272 has a curved section 272 where the capstan rollers 248B, 248C The capstan 248 is curved to compensate for the kinetic force applied to the peeled bill. It roughly corresponds to the surrounding area.

In order to guide the bill propelled by the capstan 248 into the curved guideway 270, A pair of idler rollers 262A, 262B are provided downstream of the take-out roller. ing. More specifically, these idlers These idler arms are attached to the idler arms 264A and 264B. This idler shaft is mounted so as to move in an arc around the shaft 266. The shaft is supported across the side walls of the device. Idler arm is selected A downward force can be exerted on the peeled banknote through an idler roller. It is spring-loaded on the idler shaft to allow the banknotes to The bill and the capstan 248 until guided into the curved section 272 of the guideway 270. Ensure continuous contact between.

The banknote transport path has an output path for banknotes downstream of the curved section 272. This output path is formed in the form of a flat section 274 and supported by idler rollers 262A, 262B. A banknote guided along the curved guide path 270 along this flat section, The banknotes are moved in the opposite direction to the direction in which they are exited from the input bin. Take-out roller 2 52A, 252B and caps provided by capstan rollers 248B, 248C. along the direction of rotation of the stun and the guidance provided by the section 272 of the curved guideway 270. The movement of the banknotes entered is as best shown by arrow 272B in FIG. Initial movement of bin 227 along fx slope 229 (see arrow 247B in FIG. ) to a direction along the flat section 274 of the output path.

Thus, a note torn from the stack of notes in the input bin is 'I! @ are the take-out rollers 252A, 252B and the capster 70-ra 248B, 248. It moves along the input path while receiving active contact with C. Then the banknotes , curved guide in positive contact with idler rollers 262A, 262B. It is guided through channel 270 onto a flat section 274 of the output channel.

In the output path, the banknotes are pushed aggressively along the flat section 274 by means of a transport roller system. will be guided. The transport roller device is positively driven axially spaced It has a plurality of transfer rollers 2g2A, 284A, 286A, and these rollers is located on a transfer shaft 287 supported between the side walls of the device. Flat The sections are provided with openings through which at least one of the transport rollers is inserted. Two, specifically rollers 282A, 284A protrude and correspondingly rotate freely. It is in counter-rotating contact with the driven rollers 292A and 294A. Passive rollers are used in the output path. Mounted on a support shaft 295 supported between the side walls of the device below the flat section 274 I'm being kicked. Passive transfer rollers 292A, 294A are active transfer rollers 282A.

284A, 286A, with the contact points facing each other. The active roller and the passive roller are placed in a flat state due to active contact. so that the banknote is in the same plane as the output path so that it can be moved along the output path while has been done. Active transport rollers 282B, 284B, 286B and these rollers A similar set of spring-loaded passive transfer rollers 292B, 294B facing the slightly less than the length of the narrow dimension of the banknote to be identified downstream of the first set of transfer rollers. It is located a short distance away. Furthermore, idler rollers 262A, 2 62B and the first set of transfer rollers along the curved guide path 259. The guided bill is connected to the idler rollers 262A, 262B and the capstan. Just before moving away from active contact with the 248 If the contact between the driving roller and the driven roller is selected to be Ru.

The active transport roller is driven at a significantly higher speed than the capstan roller. passive lo The first roller of the transfer roller rotates freely and the active roller is actively driven. The banknotes fed between the rollers along the first section of the output path are driven by the active controller. It is drawn into the 27-hole formed between the roller and the passive opening. active transfer roller The high speed of causes a sudden acceleration on the banknote. This acceleration is caused by the paper being acted on by the transport rollers. Immediate separation of the banknote from any other banknote guided in the curved guideway with the banknote It functions as if it were to be peeled off.

Downstream of the first set of transfer rollers, the banknote is moved along the flat section to the active transfer rollers of Wc2. roller and a set of passive transfer rollers. second active The rollers of the set of transfer rollers and passive transfer rollers move at the same speed as the first set of transfer rollers. Rotated in degrees. Preferably, pairs of opposed active transfer rollers 282A-282 B, 284A-284B, and 286A-286B are related to each other at a pace of 290 , so that the passive rotational action of the transfer shaft 287 causes the second transfer/yaft 288 Applied to a roller supported above. A second set of transfer rollers is arranged along the output path. The active contact exerted by the rollers on the moving bill ensures that the bill is received between the first pair of transfer rollers. Arranged to occur before being released from dynamic contact! has been done. Thus, the second transfer A set of feed rollers passively guides the banknotes onto the stacker platform 235; Stacker wheels 238 and 240 pick up banknotes in this staff platform. and place it on the stacker plate 242.

Next, with particular reference to Figures @14 and 15, these figures include Figures 11 and 1. A side view and a plan view of the livestock processing equipment shown in Figures 2 and 13 are shown, respectively.

Each of these side and top views shows how the banknotes are transported into three sections of the transport path, namely the input path. along the curved guideway, and along the output path. Mechanical arrangements for driving the various means for transporting are shown. this As shown in these figures, motor 300 captures rotational motion with a belt/pulley arrangement. Used to add to stun shaft 249. The belt/pulley device is It has a pulley 310 provided on the stand shaft 249, and the pulley 310 is The drive shaft 7H of the motor is connected via a pulley 304 and a belt 306 provided. There is. The diameter of the drive pulley 310 is determined from the typical high speeds at which the motor 300 operates. In order to obtain the desired deceleration, the motor pulley 304 is selected to be suitably larger than the motor pulley 304. There is.

The drive shaft 247 for the drive roller 246 has a drive shaft mounted on the drive shaft. Rotary motion is applied by the pulley 308, and the pulley 308 9; connected via a belt 312 to a corresponding pulley 310; .

Pulleys 308 and 310 are of the same diameter so that the shaft of the drive roller The shaft 247, and therefore the drive roller 246, rests on the capstan shaft 249. It rotates in unison with the capstan 248 provided.

A transfer roller corresponding to the first set of transfer rollers is used to apply rotational motion to the transfer rollers. A pulley 314 is attached to the race seat 7t 287, and this pulley 314 is connected to the belt 31. associated with the corresponding pulley 316 of the capstan shaft 249 via B . The diameter of the transfer roller pulley 314 determines the speed from the capstan roller to the transfer roller. The corresponding capstan pulley 31 is designed to achieve step-by-step speed up to The diameter is selected to be suitably smaller than the diameter of 6. Transfer roller shaft 2 A second set of transfer rollers attached to 88 connects to the transfer pulley via belt 322. A pulley 320 associated with 314 is driven at the same speed as the rollers of the first set of transfer rollers. be moved.

As shown in FIGS. 14 and 15, the optical encoder 299 is an optical encoder according to the present invention. Details I above in connection with the detection correlation technique; Accurate tracking of the lateral movement of held banknotes with respect to the rotational movement of the transport nosect one of the transfer roller shafts, preferably passively driven, for It is attached to the conveyor belt 288.

Intermediate pulley 322 is mounted on a suitable support to drive stacker wheels 238,240. Mounted on a holding means (not shown), the boot is attached to the capstan gloss 7 to 249. It is associated via a belt 326 with a corresponding pulley 324 provided above.

Transfer of banknotes torn from the stack of banknotes in the input bin with three compartments Due to the time required to transfer the processed material through the The speed at which the stacker wheel can rotate to deliver the number of banknotes to the stacker plate must be Naturally, it is lower than the speed of the capstan shaft. Therefore, the intermediate pulley 322 The diameter is smaller than the diameter of the corresponding capstan pulley 324 to achieve deceleration. selected to be large. Intermediate pulley 322 has an associated pulley 328. , the second boob is on the drive shaft 241 for the stacker wheels 23g and 240. A stacker pulley 330 and a belt 332 are associated therewith. Figure 11~ In the preferred embodiment shown in Figure 15, the stacker wheels 238, 240 are Rotates in the same direction as the stun roller. This is the belt between pulleys 32B and 330. 332 with a fixed pin 333 placed between these two pulleys. This is done by configuring it in a figure 8 J configuration.

The curved section 272 of the guideway 270 allows conventional techniques such as double detection, length detection, skew detection, etc. Issuance for performing standard banknote handling operations such as counterfeit banknote detection using technology. An optical sensor device 299 including a light diode (LED) 298 is provided on its underside. There is. However, unlike conventional devices, banknote identification according to face value is This is not done in this area for reasons discussed below.

According to a feature of the invention, optical detection of banknotes by means of the improved optical detection correlation technique described above is provided. The scan is arranged along a flat section 274 of the output path downstream of the curved guide path 270. This is done with an optical scanning head 296. More specifically, the optical head 296 It is located between two sets of transfer rollers below the flat section of the output path. Benefits of this method The point is transferred between two sets of transfer rollers at each end of the banknote along the narrow dimension of the banknote. Optical scanning occurs on the banknote when the banknote remains approximately flat as a result of contact. This means that it will be carried out.

It will be appreciated that the drive devices described above are mentioned for exemplary purposes. three applying the rotational motion necessary to cause the movement of the banknote along a transport path with a section of Alternate configurations may be used to advantage as well. However, the most To achieve proper banknote separation, the surface velocity of the banknote across the two sets of transport rollers is The speed must be greater than the speed at which the banknote crosses the capstan roller. is important. sudden acceleration of the banknote once it contacts the first set of transfer rollers; It is this speed difference that causes the

The drive device may have a one-way clutch (not shown), which one-way clutch Provided on the capstan shaft, capstan shaft, transfer roller shaft , and the stacker wheel shaft is provided with a flywheel device (not shown). It is better if it is The combination of one-way clutch and flywheel is used after banknote identification. The banknotes remaining on the transfer path are moved from the transfer path to the stacker by the inertia of the flywheel device. Rapid batch processing of banknotes by allowing them to be automatically drawn into plates It can be used advantageously in carrying out.

As mentioned above, implementation of the optical detection correlation technology of the present invention can be applied to several denominations of currency. Only a relatively small number of reflection samples are required to properly differentiate between the two.

This allows very accurate identification even when the banknote is scanned along its narrow dimension. It can be carried out. However, the accuracy of identifying the face value is limited by the test pattern. The correlation between the reflection samples of the and the corresponding samples of the stored master pattern It is based on the degree of kinship. Therefore, the banknote is transported flat across the identification means. It is important that the That is.

In the banknote handling device shown in FIGS. 11 to 15, two sets of transfer rollers are used. Positioning an optical scan head 296 on one side of the flat section 274 of the output path between the It is done by two. In this region, the banknote was in passive contact with two sets of rollers. the banknote is held in a substantially flat position across the scanning head. Make sure to move in the same manner. Additionally, a second set of passive transport rollers is enabled in this region. At the same speed, the dynamic transfer rollers and the belt connecting these two sets of rollers Since the banknotes are driven, a uniform movement speed of the banknotes is maintained. Rad 296 to optical scanning By placing the flat section 274 downstream of the curved guideway, the identification Reflection samples obtained by optically scanning the banknotes to be separated and Maintain direct correspondence between corresponding samples of stored master patterns .

According to a preferred embodiment, the optical scanning head is located on the transport path below the scanning head. combined to evenly illuminate a light strip of the desired size onto a defined banknote. It has multiple light sources acting on it. As shown in FIG. Rays 340A and 3 are placed on the flat section 274 of the output path in which the scanning head is positioned. It has a pair of LEDs 340, 342 each pointing downwardly at 40B. LED3 40.342, whose respective rays combine to form the desired light strip 342. It is arranged at an angle to the vertical axis Y so as to illuminate.

A scanning head 296 scans the true side of the strip to detect the light reflected by the strip. Center I on top: Arranged! ; has a photoelectric detector 346; The photoelectric detector 346 detects a central processing unit ( (CPU) (not shown). Preferably, the desired dimensions are irradiated! varnish To realize the trip, rays 340A and 340 from LmD 34o, 342 B passes through the optical mask 343, respectively.

To capture reflection samples with high accuracy, a photoelectric detector is illuminated with reflection data. It is important to capture evenly across the strip. In other words, photoelectric detector 3 46 is centered above the light strip relative to the center point "0" of the light strip 17, the output of the photoelectric detector is It must optimally approximate a step function as a function of distance from the center point ``0'' along the axis. Must be. When using a nest light source placed at an angle to the vertical direction, In this case, the variation in the output of the photoelectric detector will typically be as shown by curve B in Figure 17. This Gaussian function i2 approaches.

According to a preferred embodiment, the two LEDs are arranged at angles a and β respectively with respect to the vertical axis. are arranged at an angle. The angles α and β are such that the resulting output of the photoelectric detector is It is selected to be as close as possible to the optimal distribution curve A in FIG. preferred practice According to the example, angles σ and β are chosen to be 19.9″ each. The output distribution of the photoelectric detector expressed by This is shown as a curve. This curve effectively unites the individual Gaussian distributions of the light sources and optimally -a Provides a composite distribution that closely approximates lA.

A light mask allows an optical scanning head to create multiple light strips of various dimensions. Fig. 18 shows the method. As shown in this figure, the optical mask 350 D. Allow light from the light source to pass through to illuminate a light strip of desired dimensions. An entirely opaque area 352 is formed with two slits 354 and 356. have More specifically, the slit 354 has a characteristic pattern on the test banknote. It has a length and a radius of approximately 1,27111 (0,050 inches). Second slit 35 6 detects the thin border line surrounding the print on the banknote as detailed in the text. It began to produce a relatively narrow irradiated slit that was used on two sides. There is. According to an exemplary embodiment, the narrow slit 356 is approximately 7.62s+a (0,3 00 inches) and a width of approximately 0.254 mm (0.010 inches).

It is clear that precise machining is required to form the slits accurately. That's it. In fact, it is difficult to process a narrow slit 356 on the optical mask 350. be. According to the preferred embodiment, this problem is solved by dividing the mask 350 into separate sections 36. 0 and 362 forms. Division 360 is one The edges are machined to correspond to the desired half section 356A of the slit 356. Ru. The second clamp 362 has a corresponding edge with the other half 356B of the slit 356. It has been processed to match. The two sections 360 and 362 are mechanically torqued together. In this connection, these sections effectively constitute a narrow slit 356. this An advantage of the method is that the two halves 356A, 35 which together form a slit 356 This means that 6B can be formed accurately. This is the process applied to the edge of the mask. , because it can be handled with much better precision than processing inside the mask itself. It is.

Engraving (contents (ζ unchanged) Toad Engraving (1 content unchanged) Engraving (no changes to the content) FIG. 4B Engraving (changes in content) Engraving (no changes in content) Engraving (no changes to the content) day ΩJji No changes to the contents of Book A) Distance between point 0 force N et al. abstract A method and apparatus for distinguishing between banknotes (17) of various denominations of the reflective features of the banknote obtained by scanning the banknote along irradiation (22)L. Optical detection based detection (18) - uses correlation techniques. A series of detected reflections The signal is applied to the illuminated strip with its narrow dimension parallel to the direction of banknote transport. reflections at a plurality of predetermined sample points as the bill is moved across the top (24). sample and digitally process the emitted light under microprocessor control. obtained by The sample data undergoes digital processing including a normalization process. This allows the reflection data to develop a unique pattern of characteristics for a given denomination. , with sufficient differentiation between characteristic patterns to distinguish between various denominations. Incorporate features.

Procedural amendment September 1992 Horn Day

Claims (16)

[Claims] 1. In a method for discriminating between banknotes of various denominations, a coherent Determination of the banknote by focusing at least one strip of light onto the banknote irradiating a section of the banknote and detecting light reflected from the irradiated section of the banknote. and outputting an analog reflected signal; illuminating, i.e. optically, successive sections of said banknote along a predetermined dimension of said banknote; a scanning pattern between the strip of coherent light and the banknote; performing opposite lateral movement; a series of analog reflections corresponding to the light reflected from each of said successive sections of banknotes; a step of obtaining a signal; The series of analog reflection signals are digitized, processed and combined to determine the banknote. The process of creating a set of digital data samples representing the data pattern characteristics of face value. With mode, a set of optical scans corresponding to the original banknotes for each of the various cheek denominations to be identified; the process of generating and storing a master feature pattern and the features of the scanned banknote; Compare the pattern with each of the stored master patterns to compare these patterns. determining the degree of correlation between and thereby identifying the face value of the banknote; How to have. 2. the relative lateral direction between the coherent light strip and the banknote; The movement holds the light strip stationary while at the same time pushing the bill into this strip. , so that the series of reflected signals is The banknote identification method according to claim 1, wherein the banknote identification method is obtained over the predetermined dimensions of the banknote. 3. Each bill has a front side and a back side, and each bill moves across said strip. The optical scanning of the banknote may be performed in a "forward" direction or is carried out along either direction of the "backward" direction, and furthermore, the four master feature patterns The pattern is stored for each denomination for which a loan is to be generated and identified; , a light beam is scanned along the front and rear directions of the banknote over the front and back sides. The patterns according to claim 2 each correspond to a pattern generated by being chemically scanned. Banknote identification method. 4. The banknotes have a wide dimension and a narrow dimension, and the reflected signal causes each banknote to have a wide dimension and a narrow dimension. Banknote identification according to claim 1, obtained by optically scanning along the dimension. Another way. 5. When scanning each bill along its narrow dimension, at least 64 analog 5. The banknote identification method according to claim 4, wherein an emission signal is obtained. 6. The banknote has a printed mark on the front and back sides, and the mark is made of paper. It is surrounded by a border line constructed on the banknote, so that the side of the banknote outside the border line is substantially blank and the series of analog signals representing the characteristic pattern of the banknote is Banknote identification according to claim 5, wherein the number is obtained from the side of the banknote enclosed within the border line. Another way. 7. A first relatively narrow strip of coherent light indicates that the banknote is The reversal obtained from outside the border line on the face of the banknote as it moves across the lip. detecting the difference in magnitude between the emitted signal and the reflected signal obtained for the edge decoration line itself; A second beam of coherent light is used to detect the border line by A relatively wide strip is formed by forming the feature pattern after the border line is detected. 7. A banknote identification method according to claim 6, which is used to obtain the reflected signal representing a banknote. 8. The characteristic pattern of the scanned banknote and the stored master pattern Following the step of determining the degree of correlation between each, the scanned banknotes are It can be seen that the degree of correlation is the highest, and the notation that is at least equal to the predetermined correlation threshold The banknote is clearly identified as having a face value corresponding to the memorized master pattern. The banknote identification method according to claim 1. 9. scanned banknotes that were not clearly identified as having a specific face value. The banknote identification method according to claim 8, wherein the banknote is identified as having an unidentifiable face value. Law. 10. Identified in devices for identifying and counting banknotes of various denominations an input path for receiving banknotes, along which the banknotes are directed in a first direction; an input path along which the banknote can be moved; an output path along which the banknote can be moved; arranged between an input path and an output path, for receiving banknotes from said input path and forwarding them; a curved guide path for guiding the bill onto the output path along the second direction; and a curved guide path for guiding the banknote onto the output path along the second direction. downstream of the curved guide path along the output path in which the output path is guided substantially straight. A banknote recognition device having a banknote recognition means placed thereon. 11. The banknote identification means includes: A predetermined section of the banknote is illuminated with at least one strip of coherent light thereon. means for irradiating the bill by assembling it across said light strip; The banknotes generate a series of analog reflection signals as they cut and move. means for detecting light reflected from the illuminated section at predetermined time intervals; The reflected signals are digitized, processed and combined to create a data of the face value of the banknote. means for creating a set of digital samples representative of the pattern characteristics to be identified; A set of master feature patterns corresponding to an optical scan of each original banknote of face value means for generating and storing a characteristic pattern of the scanned banknote; Determine the degree of correlation between these patterns by comparing them with each of the master patterns and means for thereby identifying the face value of the banknote. Item 10. The banknote identification device according to item 10. 12. The banknote has a wide dimension and a narrow dimension, and the banknote has a wide dimension and a narrow dimension. the curved guide along the input path while remaining substantially parallel to the input path; The banknote identification according to claim 10, which is moved along the inner path and along the output path. Separate device. 13. The banknotes have a front side and a backside, and each banknote moves across the strip. By doing so, the optical scanning can be performed in either the “forward” direction or the “backward” direction. furthermore, for each denomination to be identified, four are stored in said storage means. four master feature patterns are stored, and the four banknote patterns have the face value The front and back sides of the original banknote of the same denomination are scanned in the forward and backward directions. according to claim 11, each generated by optically scanning along a direction. Banknote identification device. 14. The banknote bears a printed indicia on the surface and characterizing the face value of the banknote. on the back side, and the mark is surrounded by a border line formed on the face of the banknote. The side of the banknote outside the border line is substantially blank, and the banknote identification The means obtains the reflected signal from a surface of the banknote surrounded by the border line. 14. The banknote identification device according to claim 13, wherein the characteristic pattern is generated by: 15. The banknote identification means detects banknotes as they move across the strip. The reflected signal obtained from the outside of the border line and the signal obtained from the border line itself. In order to detect the border line by detecting the difference in magnitude between the reflected signal and the reflected signal. , for focusing a first relatively narrow strip of coherent light onto the banknote. 15. The banknote identification device according to claim 14, further comprising means for identifying the banknote. 16. The banknote identification means uses the narrow strip of light to identify the border line. After the characteristic pattern is detected, a coherent lens is used to obtain the reflected signal representing the characteristic pattern. means for collecting a second relatively wide strip of light on said banknote; The banknote identification device according to claim 15.