JPH0547054B2 - - Google Patents

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention is applicable to the verification of authenticity and defacement of objects in banknote verification devices, etc., and the detection of overlapping objects in paper feeding devices that feed paper sheets one by one. The present invention relates to a variable gain adjustment device for adjusting the output level of an optical sensor used for detecting the position and direction of objects, identifying defects, etc. in industrial robots on assembly lines.

(Prior Art) Conventionally, in a banknote authentication device using this type of automatic gain adjustment device, an optical sensor is generally used as an authentication sensor, and the target banknote to be validated is irradiated with light emitted from a light source. The reflected or transmitted light is photoelectrically converted by a light-receiving element, and features are extracted from the pattern of strength and weakness of the resulting electrical signal for identification.

FIG. 10 shows an example of the configuration of an optical sensor used for appraisal. Figure a shows an example of the configuration of a reflective optical sensor. The sensor case 6a, which has a substantially L-shaped cross section, is provided with a light emitting element 1 at one end (top), a cover glass 3 facing an object 5 such as a banknote at the other end (bottom), and a middle part of the sensor case 6a. A light receiving element 2 is provided opposite the slit 4 and the cover glass 3. In this optical sensor, after light emitted by a light emitting element 1 is restricted from passing through a slit 4, a predetermined area of a target 5 is illuminated through a cover glass 3, and the reflected light is received by a light receiving element 2. to obtain an electrical signal.

FIG. 1B shows an example of the configuration of a transmission type optical sensor. One long sensor case 6b is provided with a light emitting element 1 at one end and a diffuser plate 7 in addition to the cover glass 3 at the other end. For this sensor case 6b, the object 5
The other short sensor case 6c, which is attached movably and facing each other, is provided with a cover glass 3 at one end and a light receiving element 2 at the other end.
In this optical sensor, light emitted by a light emitting element 1 is converted into diffused light with uniform light density by a diffusion plate 7, and then irradiated onto an object 5 through a cover glass 7, and the light transmitted through the object 5 is transmitted to a light receiving element. 2 to receive the light and obtain an electrical signal.

Figure c shows another configuration example of a reflective optical sensor, in which a diffuse reflector plate 8 having a reference reflectivity is used to obtain a sufficient output level even when there is no object 5, compared to the configuration example of the reflective optical sensor shown in figure a. is provided on the outside of the sensor case 6a facing the cover glass 3.

In this way, the output level of the optical sensor directly obtained from the light-receiving element 2 has large gain variations due to differences in the light output of each light-emitting element 2 and differences in photoelectric conversion efficiency of each light-receiving element 2. requires extensive gain adjustment. Errors in this gain adjustment directly affect the discrimination performance, and depending on how the gain is adjusted, may cause an increase in the reject rate or erroneous identification, so it is necessary to perform the gain adjustment with high precision.

For this reason, conventionally, the gain adjustment of the output level of the optical sensor was performed manually using a variable resistor VR, but VR adjustment in a banknote validation device that has a large number of sensor pairs requires time and adjustment work. It was difficult to obtain stable adjustment accuracy.

An example of the configuration of a conventional automatic gain adjustment device that solves this problem is shown in FIG. The automatic gain adjustment device shown in FIG. By adjusting the gain so that a predetermined output is obtained in the non-intervention state, the output of the optical sensor whose gain is adjusted when the target 5 is discriminated. It's trying to get a signal.
That is, this automatic gain adjustment device includes a variable gain amplifying means 11 that amplifies (or attenuates) the output of the optical sensor based on a set gain, and compares the non-intervention level with a preset non-intervention level reference amount. and comparing means 12 for setting the gain. As the optical sensor in this case, for example, the one shown in FIG. 10c described above is used. In the operation of this device, first, a predetermined gain is set in the variable gain amplification means 11. Next, the comparison means sets a gain in the variable gain amplification means 11 that minimizes the error between the output level of the optical sensor without intervention, for example, the output level of the variable gain amplification means 11, and the reference amount of the non-intervention level. be done. As a result, an adjusted optical sensor output signal is obtained at the output of the variable gain amplification means 11.

The automatic gain adjustment device shown in FIG. 11b performs gain adjustment in accordance with the output of the optical sensor in the immediately previous (past) intervening state of the object 5.
That is, this device is comprised of variable gain amplification means 11, comparison means 13, and gain modification means 14.
The comparison means 13 compares the past intervention level, for example, the output level of the variable gain amplifier 11, with the intervention level reference amount, and outputs a correction amount that reduces these errors. The gain correction means 14 is the comparison means 1
Based on the amount of correction from 3, the gain obtained immediately before is corrected and set in the variable gain amplification means 14. As a result, the adjusted output level of the optical sensor is obtained as the output of the variable gain amplification means 11.

A case will be described in which this automatic gain adjustment device is applied to, for example, a banknote validation device.

Banknote appraisal is performed by measuring the output of the optical sensor for appraisal in the banknote intervening state for each banknote, but in gain adjustment, the measurement result of this intervening level is treated as an observation amount for adjustment error detection. First, the intervention level is statisticized for a predetermined number of banknotes, and the gain of the variable gain amplification means 11 is determined by the comparing means 13 according to the statistical amount.
and determined by the gain correction means 14. The temporarily obtained statistical value is the statistical standard value (intervention level standard value)
If the amplification rate is 10% lower than that, adjust the amplification factor to a 10% increase.

(Problems to be Solved by the Invention) However, the automatic gain adjustment device having the above configuration has the following problems.

The automatic gain adjustment device described in FIG. Variations in optical characteristics; (a) Variations in directivity for each light-receiving element; (c) Variations in the degree of diffusion for each diffuser plate (or diffuse-reflection plate); Due to factors such as variations in mounting position, the output ratio of the optical sensor between the target state and non-intervening state differs for each optical sensor, so even if the output level for the non-intervening state is precisely adjusted, the target intervening state The output level varies and adjustment accuracy cannot be obtained. Therefore, when this automatic gain adjustment device is applied to a banknote validator, such adjustment accuracy errors make strict appraisal difficult and reduce the reliability of the banknote validator.

On the other hand, in the automatic gain adjustment device described in FIG. 11b, automatic gain adjustment is performed by measuring the output level of the optical sensor in the state where the object is present, and there is a large adjustment error due to the factors (a) to (d) above. Although this does not occur, the followability of the adjustment operation deteriorates due to one of the following two causes.

(E) Since the output level of the optical sensor of the target as an observation quantity for adjustment varies from target to target, statistical processing is required for multiple targets, which lengthens the statistical observation period.

(f) For operational reasons, there are cases where there is a very long period in which there is no target for obtaining adjustment observables.

As a result, when this device is applied to, for example, a banknote authentication device in a currency exchange machine, the output level of the optical sensor of the banknote to be verified will depend on the thickness of each banknote, the darkness of the print, the misalignment of the print, and the dirt caused by circulation. Since there are large variations from banknote to banknote depending on the condition, etc., it is necessary to conduct statistics on hundreds or thousands of banknotes in order to use it as an observation amount for adjustment, and the tracking of adjustment operations is poor due to the long observation and statistical period. I get used to it. Furthermore, the observation statistics period depends on the operational status of the money exchange machine, and can be extremely long if the frequency of money exchange is low. Therefore, in general, this automatic gain adjustment device does not perform adjustment operations for fluctuations in the optical sensor output that change over a short period of time, such as changes over time in the optical output of the light source immediately after power is turned on, or changes over time in the sensor output due to the temperature inside the device. The problem was that it could not be followed.

SUMMARY OF THE INVENTION An object of the present invention is to provide an automatic gain adjustment device capable of solving the above-mentioned problems of adjustment accuracy and adjustment followability, and stably obtaining a highly consistent adjustment output.

(Means for Solving the Problems) In order to solve the above-mentioned problems, the present invention provides, in a first invention, an automatic gain adjustment device provided at the output of an optical sensor that photoelectrically converts light from an object.
(a) variable gain amplification means that amplifies or attenuates the output of the optical sensor based on a set gain; (b) an output level of the optical sensor when the target is not present and a first reference level that is set in advance; (c) an error between the past output level of the optical sensor when the target exists and a second reference level set in advance; (d) updating means for immediately updating the gain correction amount based on the error amount from the second comparison means so that the error is minimized;
and (e) a modifying means for modifying the gain of the first comparing means based on the gain modification amount from the updating means and setting it in the variable gain amplifying means.

Further, in a second invention, in an automatic gain adjustment device provided at the output of an optical sensor that photoelectrically converts light from a target, (a) a variable that amplifies or attenuates the output of the optical sensor based on a set gain; gain amplification means; (b) second comparison means for outputting an error amount between the past output level of the optical sensor when the object was present and a second reference level set in advance; (c) the second comparison means (d) an output level of the optical sensor when the target is not present; and (d) an output level of the optical sensor when the target is not present. The first comparison means sets the gain to the variable gain amplification means such that the error from the first reference level from the update means is minimized.

Furthermore, in a third invention, in an automatic gain adjustment device provided at the output of a photosensor that photoelectrically converts light from a target, (a) a variable device that amplifies or attenuates the output of the photosensor based on a set gain; gain amplification means; (b) second comparison means for outputting an error amount between the past output level of the optical sensor when the object was present and a second reference level set in advance; (c) the second comparison means (d) when the target does not exist based on the level correction amount from the updating means; (e) modifying means for modifying and outputting the output level of the optical sensor; and (e) modifying the variable gain so that the error between the output level of the modifying means and a preset first reference level is minimized. The first comparing means is provided in the amplifying means.

(Function) The technical means of the present invention functions as follows. In the first invention, the gain obtained by the first comparison means from the output level of the optical sensor when there is no target (unmediated level) is calculated from the past output level of the optical sensor when the target is present (intervening level). The gain is corrected according to the gain correction amount obtained by the second comparing means and the updating means. Therefore, by the gain correction effect based on the non-intervention level by the first comparison means, it is possible to quickly follow short-term optical sensor output fluctuations due to temperature fluctuations, etc. By correcting the ratio between the non-intervention level and the intervention level by the gain correction amount, it is possible to accurately adjust for long-term fluctuations due to dust or the like.

In the second invention, the first comparison means calculates the gain without intervention according to the second reference level (non-intervention level reference amount) obtained from the past intervention level by the second comparison means and the update means. I'm looking for. In addition, in the third invention, the second level of intervention is determined from the past intervention level.
Based on the level correction amount obtained by the comparison means and the updating means, the correction means corrects the non-intervention level, and obtains the gain to be set by the first comparison means from the corrected non-intervention level. There is. Therefore, like the first invention, both the second and third inventions have the function of correcting the ratio between the non-intervention level and the intervention level, and the function of correcting the gain based on the non-intervention level. Technological problems can be solved.

(Embodiment) FIG. 1 is a functional block diagram of an automatic gain adjustment device showing a first embodiment of the present invention. In this figure, the same reference numerals as in FIG. 11 indicate the same components. The automatic gain adjustment device of this embodiment is as follows:
Variable gain amplification means 11, comparison means 12a, 13
a, update means 15, and correction means 16. The comparison means 12a is the comparison means 1 of FIG. 11a.
Similarly to 2, the non-intervention level is compared with a preset non-intervention level reference amount, and the gain of the variable gain amplification means 11 that minimizes the error with the reference amount is output. However, the comparison means 12a of this embodiment
outputs this gain as a temporary gain (gain before gain modification). On the other hand, the comparison means 13a
Similar to the comparison means 13 in FIG. b, the past intervention level is compared with the intervention level reference amount to detect the error amount for the reference amount. output the reference amount). The updating means 15 updates the previous gain correction amount based on the correction amount which is the reciprocal of the error amount. The modifying means 16 multiplies the temporary gain from the comparing means 12a by the gain modification amount from the updating means 5 and outputs a gain.

In addition, in the figure, the input of the non-intervention level of the comparison means 12a is the input of the variable gain amplification means 11,
The input of the intervention level of the comparison means 13a is connected to the output of the variable gain amplification means 11, but this is to conveniently indicate that the non-intervention level is the current one and the intervention level is the past one (hereinafter referred to as , the same applies to the second to sixth embodiments described later).
Therefore, the non-intervening level may be the output of the variable gain amplifying means 11, or the input of the intervening level variable gain amplifying means 11. In this case, since the intervention level is in the past, a storage means is required. Furthermore, if this embodiment is implemented using a microprocessor or the like, an A/D converter or the like is required.

Next, the operation of the first embodiment will be explained.

First, an initial value gain is set in the variable gain amplification means 11. Next, the comparison means 12a measures the output level of the variable gain amplification means 11 in the non-intervention state of the object 5 as the non-intervention level, and the correction means 16 sets a gain that minimizes the error with respect to the non-intervention level reference amount. Output to. For example, a gain is output to the correction means 16 such that the output level falls within a predetermined range of output levels set in advance as a reference amount. On the other hand, the past (before the previous time)
The intervention level in the intervention state of one or more objects 5, for example, the output level of the variable gain amplification means 11 and a preset intervention level reference amount are compared by the comparison means 13a, and the ratio of the intervention level to the intervention level reference amount is determined. is output to the updating means 15 as an error amount. The updating means 15 takes the reciprocal of the ratio from the comparing means 13a, and the obtained correction amount,
The gain correction amount is updated by multiplying it by the gain correction amount obtained immediately before and is output to the correction means 16. Modifying means 16 that receives this gain modification amount
Then, by multiplying this by the gain received from the comparing means 12a, the gain is corrected and set in the variable gain amplifying means 11. As a result, the adjusted output of the optical sensor is obtained as the output of the variable gain amplification means 11.

FIG. 2 is a functional block diagram of an automatic gain adjustment device showing a second embodiment of the present invention. The difference between this embodiment and the first embodiment is the intervention level comparing means 13b and the updating means 15a. That is, in the first embodiment, the comparing means 13a outputs the ratio between the intervention level and the intervention level reference amount as the error amount to the updating means 15, and the updating means 15 multiplies the reciprocal of this ratio by the immediately preceding gain correction amount. The gain correction amount was updated by doing this. On the other hand, in this embodiment, the comparison means 13
b outputs a signal indicating the error amount of the intervention level with respect to the intervention level reference amount to the updating means 15a. The updating means 15a updates the previous gain correction amount by adding or subtracting a predetermined amount based on this signal so as to reduce the error. In other words, comparison means 1
3a compares the intervention level with the intervention level reference amount and if the intervention level is large, the updating means 15a
adds or subtracts a predetermined amount from the previous gain correction amount to make the gain smaller, and when the intervention level is small, the updating means 15a adds or subtracts a predetermined amount from the previous gain correction amount to make the gain larger. By adding and subtracting, a new gain correction amount is obtained.

FIG. 3 is a functional block diagram of an automatic gain adjustment device showing a third embodiment of the present invention. In this figure, the same reference numerals as in FIG. 1 indicate the same components. The device of this embodiment includes a variable gain amplifying means 11, comparing means 12b and 13a, and an updating means 15.
Consists of b. The updating means 15b is the comparing means 1
Based on the error amount (ratio) from 3a, the previous non-intervention level reference amount is updated so that the error is minimized. The comparison means 12b sets the gain of the variable gain amplification means 11 so that the error of the non-intervention level with respect to the updated non-intervention level reference amount becomes small.

Next, the operation of the third embodiment will be explained. first,
The comparing means 13a calculates the ratio between the intervention level in the past intervention state of the target and the preset intervention level reference amount, and outputs it to the updating means 15b. The updating means 15b multiplies the reciprocal of the ratio from the comparing means 13a by the previous non-intervention level reference amount and outputs the updated non-intervention level reference amount to the comparing means 12b. Then, the comparison means 12b calculates a gain that minimizes the error of the non-intervention level with respect to the non-intervention level reference amount from the updating means 15b, and this gain is such that the variable gain amplification means 11
is sent to and configured. As a result, the output of the optical sensor is based on the set gain of the variable gain amplification means 1.
1 to obtain a regulated output.

FIG. 4 is a functional block diagram of an automatic gain adjustment device showing a fourth embodiment of the present invention. The difference from the third embodiment is the intervention level comparing means 13b and the non-intervention level reference amount updating means 15c. The updating means 15c updates the non-intervention level reference amount by adding or subtracting a predetermined amount based on the signal indicating the error amount from the comparing means 13b, which is the same as the updating means 15 of the second embodiment, so that the error is minimized. do.

FIG. 5 is a functional block diagram of an automatic gain adjustment device showing a fifth embodiment of the present invention. In this figure, the same reference numerals as in FIG. 1 indicate the same components. The device of this embodiment includes variable gain amplification means 11, comparison means 12c and 13a, and updating means 15.
d. It is composed of a correction means 17. Update means 15
d updates the previous non-intervention level correction amount based on the error amount (ratio) from the comparing means 13a so as to minimize the error. The correction means 17 multiplies the non-intervention level by the updated non-intervention level correction amount to correct the non-intervention level. The comparison means 12c sets a gain in the variable gain amplification means 11 that minimizes the error of the corrected non-intervention level with respect to a preset non-intervention level reference amount.

Next, the operation of the fifth embodiment will be explained.

First, the comparing means 13a calculates the ratio between the past intervention level and a preset intervention level reference amount, and outputs the ratio to the updating means 15d. Update means 1
5d, the product of the reciprocal of the ratio from the comparing means 13a and the immediately preceding non-intervention level correction amount is output to the correction means 17 as the updated non-intervention level correction amount. Thereafter, the non-intervention level is multiplied by the updated non-intervention level correction amount by the correction means 17,
The corrected unmediated level of this result is Comparison Means 1
Sent to 2c. And, in the comparison means 12c,
A gain that minimizes the error between the non-intervention level sent from the correction means 12c and a preset non-intervention level reference amount is sent to the variable gain amplification means 11 and set. As a result, the adjusted output of the optical sensor is obtained as the output of the variable gain amplification means 11.

FIG. 6 is a functional block diagram of an automatic gain adjustment device showing a sixth embodiment of the present invention. The difference between this embodiment and the fifth embodiment is the comparison means 13b for the intervention level and the updating means 15e for updating the correction amount of the non-intervention level. That is, the updating means 15e updates the position so that the error is minimized with respect to the previous non-intervention level correction amount, based on the signal indicating the error amount from the comparison means 13b, which is the same as in the second and fourth embodiments. Update the non-intervention level correction amount by adding or subtracting the quantitative amount.

Next, an application example in which the automatic gain adjustment device according to the present invention is applied to a banknote validation device will be described.

FIG. 7 is a block diagram showing an example of the configuration of a banknote validation device to which the automatic gain adjustment device of the first embodiment is applied. The banknote validation device shown in FIG. Converter 24, latch circuit 15a, 1
5b, interface (I/F) circuit 26,
It consists of a ROM 27, a nonvolatile RAM 28, a RAM 29, and a microprocessor 30 that controls them, and these components are connected via an internal bus. An electrical signal obtained by the light receiving element 2 of a reflective or transmissive optical sensor 20 used for banknote authentication is amplified by a corresponding amplifier 21. Based on the selection number n of the photosensors 20 given by the microprocessor 30, the analog multiplexer 22 selects only one photosensor among the plurality of photosensors 20, and controls the amplifier 21 corresponding to that photosensor. Output the output signal. The output of analog multiplexer 22 is input to variable gain amplifier 23. On the other hand, the microprocessor 30 sets the gain Gn for the variable gain amplifier 23. The variable gain amplifier 23 multiplies the output of the optical sensor amplifier 21 by Gn and outputs the result to the A/D converter 24. The A/D converter 24 converts the output signal of the optical sensor amplifier 21 multiplied by Gn into an A/D converter.
The conversion result is sent to the microprocessor 27.
Read by An interface (I/F) circuit 26 for the higher-level control device notifies the microprocessor 30 when receiving various commands from the higher-level control device, and also notifies the microprocessor 30 of the judgment result after executing the command. Notifies the higher-level control equipment of other command execution results. ROM2
In addition to the control means (program) for the microprocessor 30, identification reference data and the like are written in the memory 7. The nonvolatile RAM 28 stores parameters K _o , Y _o , and N for adjusting the output of the optical sensor, which will be described later, and retains the data even when the power of the device is turned off. Here, the letters with the subscript n attached to the English letters (K _o , Y _o, etc.) are originally multiple pieces of data independent of each optical sensor 20, but the representative value is data regarding the optical sensor with sensor number n. At the same time, it indicates that the same applies to optical sensors with other sensor numbers.

Next, the operation of the banknote validation device will be explained.

First, for example, when power is turned on for the first time to a completed banknote validation device in a factory, the parameters K _o ,
Since Y _o , N, and G _o are not stored in the non-volatile RAM 28, the host controller issues a sensor adjustment initialization command to cause the microprocessor 30 to perform the following initial settings. .

That is, the microprocessor 30 first calculates the gain modification ratio (gain modification amount) K _o and the statistic (intervention level).
Let Y _o and the statistical number N of sheets be K _o =1, 0 Y _o =0 N=0, respectively. Next, the microprocessor 30 calculates the gain
Go through the _initial setup steps. This procedure is the same as the adjustment process when there is no intervention during banknote appraisal, which will be described later, and a flowchart thereof is shown in FIG. By executing this adjustment process without intervention, the microprocessor 30
It plays the role of the comparison means 12a described in the embodiment.
The initial setting procedure for the gain G _o (adjustment process without intervention) will be described below with reference to FIG.

First, the gain G _o of the variable gain amplifier 23 is temporarily set as g ₀ (S1). Here, g ₀ is determined at the time of design, but it can be arbitrarily determined as long as g _nax g ₀ g _nio (determination range in step S5). Next, the output of the A/D converter 24 when the gain G _o (=g _o ) is examined. When the output X _o of the A/D converter 24 is within the predetermined output range (X _nax X _o X _nio ), the currently set gain G _o becomes the adjustment result, and the adjustment process ends. This predetermined output range corresponds to the non-intervention level reference amount of the first embodiment.
On the other hand, if the output X _o of the A/D converter 24 is larger than the predetermined output range (X _o > X _nax ), the gain G _o is decreased, and on the other hand, the output is smaller than the predetermined output range. (when X _o < X _nio ), increase the gain G _o (S3, S
4). After that, as a result of gain increase/decrease, the current gain G _o obtained is within the designed range (g _nax G _o
If the amplification factor is less than the lower limit value _{g nio ,} the amplification factor (gain) of the variable gain amplifier 23 is set to the lowest value. A/D converter 2
Since the output of 4 is large, it is determined that there is an excessive sensor output error, and conversely, if the gain exceeds the upper limit g _nax , it is determined that there is an under-sensor output error (S6, S
7). In either of these cases, the content of the error determination is notified to the higher-level control device, and the adjustment operation is terminated. If the current gain G _o is within the designed range (g _nax G _o g _nio ), then
Continue the adjustment operation and repeat from step S2.
In addition, at the time of design, g _nax , g _nio , and X _nax ,
The value _of
The resolution of the D converter 24, as well as X _nax and X _nio , are properly determined based on the ultimately required accuracy specifications.

In this way, by repeating steps S2 to S5 and correcting the value of gain G _o , if there is no abnormality in the sensor output, the A/D conversion output will change to the predetermined output X _nax X _o within a limited number of repetitions. Find the gain G _o such that X _nio ). The obtained value of G _o is set as the initial value of G _o by the microprocessor 30.
Write to RAM28 and complete the initial setting procedure of _Go .

Next, automatic adjustment by the microprocessor 30 during banknote appraisal and associated processing will be described with reference to the flowchart of FIG.

First, when the microprocessor 30 receives a command to authenticate one banknote from the host control device via the I/F circuit 26, it determines whether or not the target banknote is in the non-intervening state (S11). . As a result of this determination, if there is no intervention, error processing is performed; on the other hand, if there is no intervention, the output of the A/D converter 24 is checked and the gain of the variable gain amplifier 23 is adjusted so that a predetermined output is obtained. G _o is determined according to the processing procedure shown in Fig. 8, and the result is stored in the non-volatile RAM 28.
(S13).

Next, the gain correction ratio K _o is read from the nonvolatile RAM 13 and multiplied by the gain G _o obtained in step S13, and the multiplication result G _o '=G _o ×K is set in the variable gain amplifier 23. The gain adjustment of the output signal of the optical sensor 20 immediately before the appraisal operation is completed.

Next, the output of the optical sensor is checked, and when the presence of the banknote to be verified is detected, the following verification operation is started.

First, each optical sensor 2
The outputs of the A/D converters 24 at various locations on the banknotes are read for each zero and stored in the RAM 29 as pattern data (S16). After obtaining characteristic data of the banknote from the pattern data of the banknote to be appraised thus obtained, the ROM
Appraisal is performed by comparing with standard data for appraisal in 27 (S17). This appraisal result is notified to the host control device (S18).

In this manner, when the appraisal process is completed, it is determined whether or not the pattern data of the appraisal result is to be processed as a measurement value (intervention level) for adjustment (S1
9). If the banknote to be appraised is not recognized as a genuine note or is found to be defaced, the execution of the appraisal command will be completed without doing anything. If so, the integrated value _So of the pattern data for each optical sensor is determined (S20). The determined integral value _So of the pattern data is added to the sum of the integral values _So of the pattern data up to now, that is, the statistical quantity Y _o (S21). This statistic Y _o
The value of the statistical number N is incremented each time the number N is added to, that is, each time a normal ticket is evaluated (S22). As a result,
When the m-th sheet (for example, m = 1000 sheets) is reached, statistical results Y _o for m sheets of pattern data are obtained, and a predetermined statistical result standard value (intervention level standard amount) Y _o and the statistical results are combined. , that is, the gain correction ratio K _o =y _o /Y _o is determined and stored in the nonvolatile RAM 28 (S24). Therefore, the value of the gain modification ratio K _o is updated. Then, after returning the statistical amount Y _o of the pattern data integral value and the statistical number N of bills to zero so that statistics can be performed again, the execution of the single banknote appraisal command is completed (S25). Incidentally, if the statistical number N of sheets is less than m at step S23, the process ends immediately.

In the above application example, as shown in Figure 7a,
Although a variable gain amplifier was provided between the output of the optical sensor 20 and the input of the A/D converter,
In the configuration shown in FIG. 7b, the variable gain amplifier includes a photosensor 20, and by setting the light amount (lighting current) of the light emitting element 1 based on a given gain G _o , the photocurrent of the light receiving element 2 is set. It is configured to amplify or attenuate the output itself.

In this way, the variable gain amplification means can be realized by controlling the light amount of the light receiving element, and it is also possible to realize the variable gain amplification means without necessarily providing it with hardware.
It is clear that the present invention can also be realized by software that multiplies and divides the data read by the A/D converter using the microprocessor 30.

Further, it is obvious that the second to sixth embodiments can be similarly applied to a bill validating device or the like.

As described above, according to the first to sixth embodiments,
By correcting the ratio between the measured non-intervening level and the intervening level and the gain correction action based on the non-intervening level, the conventional problems such as changes in light output over time after the light source is turned on and the temperature inside the device can be resolved. Gain adjustment can be performed with good followability even for short-term output fluctuations such as changes over time in the sensor output, and no adjustment errors occur due to variations in the ratio between the non-intervention level and the intervention level.

Therefore, it is no longer necessary to set a margin in advance to take into account the output fluctuation range of the optical sensor in the discrimination criteria at the time of discrimination, and extremely precise discrimination can be performed.
The effect of eliminating the need for maintenance and adjustment can also be expected.

(Effects of the Invention) As explained in detail above, according to the present invention,
The ratio of the output level of the optical sensor when there is no target and the output level of the optical sensor when there is an object is corrected, and the gain is set based on the output level of the optical sensor when there is no target, so it can be used for a short time. It is possible to follow fluctuations in the output of the optical sensor, and to adjust the output level of the optical sensor with high precision.

[Brief explanation of the drawing]

1 to 6 are functional block diagrams of automatic gain adjustment devices showing first to sixth embodiments of the present invention, respectively, and FIGS. 7a and 7b are banknotes to which the automatic gain adjustment device according to the present invention is applied. FIG. 8 is a flowchart showing an adjustment process without intervention; FIG. 9 is a flowchart showing an automatic adjustment process during banknote authentication; FIGS. 10a, b, and c are Cross-sectional view showing an example of the configuration of an optical sensor, FIGS. 11a and 11b
1 is a functional block diagram showing an example of the configuration of a conventional automatic gain adjustment device. 1... Light emitting element, 2... Light receiving element, 3... Cover glass, 4... Slit, 5... Target, 6a~
6c...Sensor case, 7...Diffusion plate, 8...Diffuse reflection plate, 11...Variable gain amplification means, 12a~
12c, 13a-13b... Comparison means, 15, 1
5a to 15e...updating means, 16, 17...correction means, 20...optical sensor, 21...amplifier, 22
... Analog multiplexer, 23 ... Variable gain amplifier, 24 ... A/D converter, 25a, 25b
...Latch circuit, 26...Interface (I/
F) Circuit, 27...ROM, 28...Nonvolatile
RAM, 29...RAM, 30...Microprocessor.

Claims (1)

[Claims] 1. An automatic gain adjustment device provided at the output of an optical sensor that photoelectrically converts light from a target, comprising: (a) a variable gain that amplifies or attenuates the output of the optical sensor based on a set gain; (b) first comparison means that outputs a gain that minimizes the error between the output level of the optical sensor when there is no target and a first reference level set in advance; (c) a second comparing means for outputting an error amount between a past output level of the optical sensor when the object was present and a second reference level set in advance; (d) an error from the second comparing means; (e) updating the gain from the first comparison means based on the gain correction amount from the updating means; An automatic gain adjustment device comprising: correction means for setting the variable gain amplification means. 2. The automatic gain adjustment device according to claim 1, wherein the past output level of the optical sensor is a statistical amount statistically processed for a plurality of past objects. 3. The second comparing means is the ratio of the output level of the optical sensor to the second reference level as the error amount, and the updating means updates the gain by multiplying the reciprocal of the ratio by the immediately preceding gain correction amount. Automatic gain adjustment device according to range 1 or 2. 4. The automatic gain adjustment device according to claim 1 or 2, wherein the updating means adds or subtracts a predetermined amount from the previous gain correction amount based on the amount of error so that the error is minimized. 5. In an automatic gain adjustment device provided at the output of an optical sensor that photoelectrically converts light from a target, (a) variable gain amplification means that amplifies or attenuates the output of the optical sensor based on a set gain; (b) (c) a second comparing means for outputting an error amount between a past output level of the sensor when the object exists and a second reference level set in advance; (c) an error amount from the second comparing means; (d) an output level of the optical sensor when the target is not present and a first reference level from the updating means; an automatic gain adjustment device comprising: first comparison means for setting a gain to the variable gain amplification means so that an error between 6. In an automatic gain adjustment device provided at the output of an optical sensor that photoelectrically converts light from a target, (a) variable gain amplification means that amplifies or attenuates the output of the optical sensor based on a set gain; (b) (c) a second comparing means for outputting an error amount between a past output level of the sensor when the object exists and a second reference level set in advance; (c) an error amount from the second comparing means; (d) updating the output level of the optical sensor when the target is not present based on the level correction amount from the updating means; (e) first comparison means for setting a gain in the variable gain amplification means such that an error between the output level of the correction means and a preset first reference level is minimized; , An automatic gain adjustment device comprising: