JPH06167851A - Image stabilizing device for electrophotographic device - Google Patents

Abstract

PURPOSE:To always obtain a stable-quality image and to reduce the number of the parts of a sensor so as to restrain the rise of the cost of an electrophotographic device. CONSTITUTION:A reference toner image is formed on a photosensitive body 11 based on a reference original 21, the density of the reference toner image is measured by a toner density sensor 22, and density variation from the initial state of toner density is measured by a subtracter 42, then the correction amounts of exposure voltage and electrostatic charge voltage required for obtaining the optimum quality of a copied image are obtained. A neuro computer 51 is allowed to learn processing procedure that the correction amount in accordance with the density variation is outputted by using the correction amount as instructer data. After learning, the toner density in the case the photosensitive body 11 is in an initial state is stored in initial density storage parts (for a light part and a dark part) 45 and 46. Hereafter, the density variation from the toner density in the initial state is stored in density change storage parts (for a light part and a dark part) 47 and 48. The exposure voltage and the electrostatic charge voltage are corrected according to the correction amount obtained by the neuro computer 51 based on the density variation.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image quality stabilizing device for an electrophotographic apparatus which stabilizes the quality of a formed image by controlling an electrophotographic process based on a toner image formed on a photoconductor.

[0002]

2. Description of the Related Art In an electrophotographic copying machine or the like, it is desirable that the density of a toner image formed on a sheet is stable in order to obtain a good copied image. However, in the electrophotographic process, the characteristics of the exposure lamp, the photoconductor, the developer, and the like are changed due to various factors such as temperature, humidity, dirt, abrasion, and changes in the characteristics of materials. The concentration also changes. In order to deal with such inconvenience, conventionally, there are various methods for controlling the exposure amount in the process from the exposure of the original document to the formation of the toner image, the charge amount of the photoconductor, and the like to correct the variation of the toner image density. It has been devised and implemented.

For example, conventionally, a method of controlling the charging voltage value based on a reference value obtained by optically detecting the toner image density has been implemented. In such a method, at the stage when the copy image quality of the copying machine is initially adjusted, the toner image density created with a specific control voltage value is detected by the reflection type optical sensor, and this value is used as a reference value for control. Remember.
After that, the same toner image density detection is performed if necessary,
The difference between the density and the reference value is converted into a correction amount for the charging voltage value of the photoconductor. The following correction characteristics are used for this correction.

Image forming operations in general copying machines are roughly classified into three types: exposure of an original, charging / exposure of a photoconductor, and development. FIG. 13 shows the characteristics of each block by dividing the image forming characteristics of the copying machine into three blocks, that is, document exposure, photoconductor charging / exposure, and development.

FIG. 1A shows the optical density of the original (horizontal axis).
And the common logarithm (vertical axis) of the reflected light amount from the original. When the optical density is D, the amount of light reflected from the image of the original is X, and the amount of light reflected from the background (white paper) of the original is X _m , the optical density is given by the equation (1).

D = −log (X / X _m ) ... (1) FIG. 2B shows the potential (horizontal axis) of the photoconductor after exposure and the amount of light incident on the photoconductor, that is, the amount of light reflected from the original. It shows the relationship with the common logarithm (vertical axis). Further, (c) of the same figure shows the relationship between the potential of the photoconductor (horizontal axis) and the density of the toner image (vertical axis). Furthermore, (d) in the figure
Is the optical density of the original (horizontal axis) and the density of the toner image (vertical axis)
It shows the relationship with. In the figure, (a) → (b)
By associating the respective values in the order of → (c), (d) of the same figure is obtained, and the copy image quality is determined based on this figure.

Alternatively, Japanese Patent Publication No. 61-29502 discloses that, from a signal corresponding to an image of a bright portion formed on a photosensitive member, exposure conditions for latent image formation or development for latent image visualization are performed. A technique for controlling the conditions and controlling the charging condition of the electrostatic latent image forming member from a signal corresponding to the image of the dark portion is disclosed.

Further, there has been proposed a method of measuring the amount of change in temperature / humidity in the copying machine and the change in sensitivity of the photoconductor, which are some of the factors causing characteristic changes, and obtaining the correction value of the control voltage from the result. (See "Electrophotographic Process Control Method Using Neural Network and Fuzzy Theory", IEEJ Preliminary Report 91-05-05).

This situation applies not only to copying machines but also to printers and facsimiles adopting an electrophotographic system. In the copying machine, the photoconductor is exposed by the reflected light from the original, but in the printer and the facsimile, the output is changed by changing the output of the exposure device by the laser light or the like according to the input image signal, and the exposure is different.

[0010]

However, in the above-mentioned conventional method for measuring the change in the toner image density and correcting the change in the image quality, the change in the image quality is caused by the change in the characteristics of any part of the image forming system such as exposure, charging and development. It is not possible to specify what it is. Therefore, it is not possible to integrally correct a plurality of control values, and it is not always possible to determine the best control value. Next, the reason will be described.

Normally, even if the voltage applied to the exposure lamp or the output voltage of the high-voltage power supply connected to the charger is kept constant, the temperature
The characteristics shown in FIGS. 13 (a) to 13 (c) change due to changes in humidity, wear of the photoconductor, dirt on the exposure lamp and charger, and the like.

For example, if only the exposure amount changes and the characteristic of (a) in the figure changes from the curve A shown by the solid line to the curve a shown by the broken line, the characteristic obtained in (d) of the figure becomes The curve c shown by is changed to the curve d shown by a broken line. With the characteristic of curve c, a good copy image quality can be obtained, but with the characteristic of curve d, the overall density becomes high and the copied image becomes dark.

Further, if only the photoconductor potential is changed and the characteristic of (b) in the figure changes from the curve (o) shown by the solid line to the curve (d) shown by the broken line, the characteristic obtained in (d) of the figure becomes The curve c changes to the curve d indicated by the alternate long and short dash line, and the density becomes low and the copied image becomes bright.

As described above, it is impossible to maintain good copy image quality for a long period of time only by keeping the control voltage values for exposure and charging constant.

Also, in the technique disclosed in Japanese Patent Publication No. 61-29502, since the conditions of exposure, charging and development are controlled independently from the signals for two types of images,
For example, when the charging conditions are changed, the characteristics relating to charging are improved, but the characteristics relating to exposure are deteriorated. Therefore, as described below, the exposure condition cannot be corrected sufficiently, and the resulting image quality may not be the best.

In FIG. 3D, the bright portion described in the above publication corresponds to an image formed in the portion A having a document density of about 0.5 or less, and the dark portion is a document density. It corresponds to the image created in the B section of 1.0 or more. In (d) of the figure, when only the exposure characteristic changes, the change in the bright portion (A portion) is large as shown by the curve D, and when only the charging characteristic of the photoconductor changes, as shown by the curve K. The change in the dark part (B part) becomes large. So in this case,
The relationship between the change in each condition and the toner density change is relatively clear, and the density correction performed by independently controlling each condition is effective.

However, when the exposure characteristic changes from curve A to curve A in FIG. 9A and the charging characteristic of the photoconductor changes from curve E to curve C in FIG.
As indicated by a two-dot chain line in (d) of the figure, the characteristics of exposure and the characteristics of charging are canceled out at the portion A to reduce the change in the density characteristic, and the change in the charging characteristic at the portion B is caused. The influence may be hidden by the influence of the change in the exposure characteristic.

When the correction of the toner density by the above correction method is simulated, it is indicated by "x" in FIG. In this figure, the horizontal axis represents the original density and the vertical axis represents the toner density. However, the characteristics obtained by the above simulation have a large deviation from the initial characteristics shown by the solid line, and the correction has not been performed accurately. I understand.

Further, in the method of measuring environmental changes such as temperature / humidity, it is necessary to measure temperature / humidity, photoconductor potential, etc. Therefore, it is necessary to provide a temperature sensor, a humidity sensor, a potential sensor, etc. It was For this reason, in addition to increasing the cost of the electrophotographic apparatus, a space for each sensor is required in the apparatus, which leads to downsizing of the apparatus and deterioration of maintainability. In addition to this, since it is difficult for the above-mentioned sensor to measure stains on the exposure lamp and the like, there is a disadvantage that the above method cannot deal with uncertain factors such as stains.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an image quality stabilizing device capable of highly accurately performing image quality correction with a relatively simple structure.

[0021]

In order to solve the above-mentioned problems, an image quality stabilizing device for an electrophotographic apparatus according to a first aspect of the present invention uses a light for detecting the density of a toner image formed on a photoconductor. A density detection unit such as a sensor, and a change in the density of the toner image, which has been changed by the execution of the electrophotographic process, with respect to the reference density of the toner image whose density is in the initial state detected by the density detection unit The density change is detected based on the density change detection means, the density change detected by the density change detection means, and the correction amount corresponding to the density change of the control value that affects the image quality in the electrophotographic process. Then, the processing procedure for outputting the correction amount of the control value corresponding thereto is learned in advance, and the learning process procedure is performed according to the learned processing procedure due to the newly input density change after learning. It is characterized in that a first learning correction means for outputting the predicted correction amount of the control value corresponding to the time change.

In order to solve the above problems, an image quality stabilizing device for an electrophotographic apparatus according to a second aspect of the present invention is a density detecting means such as an optical sensor for detecting the density of a toner image formed on a photoconductor. And density change detecting means for detecting a change in the density of the toner image, which has been changed by the execution of the electrophotographic process, with respect to the reference density of the toner image whose density is in the initial state detected by the density detecting means. An operating status detecting means for detecting a cumulative operating status of the electrophotographic process since the toner image density is in an initial state, a density change detected by the density change detecting means and an electronic status detected by the operating status detecting means. Based on the cumulative operation status of the photographic process, the density change of the control value that affects the image quality in the electrophotographic process, and the correction amount according to the cumulative operation status. Then, the processing procedure for outputting the correction amount of the control value corresponding to the input of the density change and the cumulative operation status is learned in advance, and the density change and the electrophotographic process accumulated newly input after learning are accumulated. A second learning correction means for predicting and outputting a correction value of the control value according to the density change and the cumulative operation status of the electrophotographic process according to the learned processing procedure according to the operation status is provided. .

[0023]

In the image quality stabilizing device for an electrophotographic apparatus according to the first aspect, when the toner density is in the initial state, the density change detecting means detects the change in the density of both toner images previously detected by the density detecting means. I'll do it. At the same time, a correction amount corresponding to the density change of the control value that affects the image quality in the electrophotographic process is obtained while operating the electrophotographic process.

Then, based on the density change and the correction amount, the first learning correction means outputs in advance a processing procedure for outputting the correction amount of the control value according to the density change when the density change is input. learn. As the control value, for example, an exposure voltage for driving the exposure lamp or a charging voltage applied to the charger is used.

After that, when the density change between the density of the toner image detected by the density detecting means and the previously detected reference density is obtained, the first learning correction means performs the processing procedure learned as described above. Therefore, the correction amount of the control value according to the density change is predicted and output.

As described above, in the above image quality stabilizing device,
Since the correction amount of the control value is obtained by the processing procedure by learning, it is possible to integrally correct a plurality of control values. Therefore, the correct control value can be corrected even on the basis of the density change including the density change due to dirt on the exposure lamp or the like.

Since the physical quantity that needs to be detected is only the density of the toner image, only the density detecting means such as an optical sensor is provided in the electrophotographic process section for detecting the physical quantity.
Therefore, it is not necessary to increase the number of sensors and accordingly secure a space for the sensors in the electrophotographic process section.

At the time of detecting the respective densities, if the toner images corresponding to the light and dark parts of the original are formed on the photoconductor, the image quality can be corrected over a wide density range.

On the other hand, the image quality stabilizing device for an electrophotographic apparatus according to claim 2 is basically the same as the image quality stabilizing apparatus for an electrophotographic apparatus according to claim 1 by the second learning correction means. The control value is corrected according to the processing procedure. Further, the cumulative operating condition of the electrophotographic process from the time when the toner image density is in the initial state is detected by the operating condition detecting means, and the result is the second value. It is used as a learning element of learning correction means. The cumulative operating status of the electrophotographic process is, for example, the number of operating times of the electrophotographic process based on the number of development times of the developing device, the number of times the exposure lamp is turned on, and the like.

If the control value is corrected by the second learning correction means in which the cumulative operating condition of the electrophotographic process is added as a learning element as described above, it is possible to predict a change in image quality due to deterioration of the developer. Therefore, the image quality can be corrected more highly.

[0031]

【Example】

[Embodiment 1] An embodiment in which the present invention is applied to a copying machine will be described below with reference to FIGS. 1 to 11.

As shown in FIG. 2, the copying machine according to this embodiment is provided with a document feeder 2 on the upper part of the main body 1.
An original placing table 3 composed of a glass table is provided below the original placing table 3, and inside the main body 1, an exposure optical system 4 is provided below the original placing table 3, and an electrophotographic process section 5 and the like are provided below. ing.

The document feeder 2 is designed to automatically feed the documents 7 set on the document tray 6 one by one to the document table 3. The exposure optical system 4 has an exposure lamp 8, a zoom lens 9, and a plurality of mirrors 10 ... The exposure optical system 4 includes a document 7 in which an exposure lamp 8 driven by an AC power source (not shown) is placed on the document table 3.
And the reflected light from the original 7 is reflected by the zoom lens 9 and the mirror 10 on the photoconductor 1 of the electrophotographic process unit 5.
It leads to 1.

The electrophotographic process section 5 includes the photoconductor 1 described above.
1, a charging device 12, a developing device 13, a transfer device 14, a fixing device 15, a cleaning device 16 and the like.

The photoconductor 11 is charged in advance by the charger 12 before the exposure, and when the light from the exposure optical section 4 forms an image, the resistance value of the surface decreases and the charge is discharged, whereby the electrostatic latent image is formed. Are formed. The developing device 13 is configured to adhere toner to the electrostatic latent image and form a toner image on the photoconductor 11 due to the difference between the potential applied inside and the potential of the photoconductor 11.

The transfer device 14 transfers the above toner image onto the paper 20 fed from the paper feed cassettes 17 to 19, and the fixing device 15 fixes the toner image on the paper 20 by heating and pressing. It is designed to let you. Further, the cleaning device 16 removes the toner remaining on the photoconductor 11 after the transfer.

On the other hand, in the vicinity of the retracted position of the exposure lamp 8 shown in the figure, a reference document 21 for forming a reference toner image at the time of toner concentration measurement is provided. Further, a toner concentration sensor 22 is provided near the paper discharge portion of the photoconductor 11 in the electrophotographic process unit 5. The toner density sensor 22 is a sensor that optically detects the density of the reference toner image formed on the photoconductor 11 based on the reference original 21.

The above-mentioned reference original 21 is formed by simply arranging plates having different densities side by side. For example, as shown in FIG. 3, two kinds of density plates 23 each having a uniform and constant density are provided.
· Step motor 2 formed in a disk shape having 24
It may be driven to rotate by 5. The reference document 21 is rotated to switch the document density.

Toner density sensor 2 as density detecting means
As shown in FIG. 4, reference numeral 2 is a reflection type optical sensor in which a light emitting portion 26 made of a light emitting diode and a light receiving portion 27 made of a Darlington type phototransistor are integrated. The toner concentration sensor 22 receives light emitted from the light emitting section 26 and reflected by the photoconductor 11 at the light receiving section 27, and outputs an electric signal according to the amount of reflected light.

Next, the control system of this copying machine will be described.

As shown in FIG. 1, in this control system, the exposure voltage is controlled by changing the voltage application time in the AC waveform of the exposure lamp 8 by the exposure lamp regulator 31, and the charger 12 outputs the high voltage power supply 32. The charging voltage is controlled by changing the voltage.

The exposure lamp regulator 31 is adapted to receive the output of the measuring voltage generator 34 or the output of the adder 35 selected by the switch 33. The measurement voltage generation unit 34 is configured to generate an exposure voltage for toner concentration measurement, and the adder 35 outputs the exposure voltage stored in the exposure voltage storage unit 36 and the neurocomputer 51 described later. The exposure voltage correction amount is added. The exposure voltage storage unit 36 is adapted to correct variations in individual copying machines and store the exposure voltage during the copying operation.

The high-voltage power supply 32 is adapted to receive the output of the measuring voltage generator 38 or the output of the adder 39 selected by the switch 37. Measuring voltage generator 38
Is configured to generate a charging voltage for toner concentration measurement, and the adder 39 uses a charging voltage stored in the charging voltage storage unit 40 and a neuro computer 51 described later.
The charging voltage correction amount output from is added. The charging voltage storage unit 40 is configured to correct variations in individual copying machines and store the charging voltage during the copying operation.

In this copying machine, as described above, an electrostatic latent image is formed on the photoconductor 11 by the light emitted from the exposure lamp 8 and reflected by the reference original 21, and the electrostatic latent image is developed. It becomes a toner image, and the density of this toner image is detected by the toner density detection sensor 22.

An A / D converter (A / D in the figure) 41 is connected to the toner concentration sensor 22, and a subtractor 42 is provided at the next stage of the A / D converter. Switches 43 and 44 are connected to the subtractor 42. The switch 43 is
The initial density storage unit 45 or the initial density storage unit 46 is selected and connected to the subtractor 42, and the switch 44 selects the density change storage unit 47 or the density change storage unit 48 and the subtractor 42.
It is designed to connect to. In addition, the density change storage unit 4
A neuro computer 51 is connected to 7.48.

The A / D converter 41 is adapted to convert the density of the toner image formed by the reference original 21 (hereinafter referred to as the reference toner image) into digital form. The initial density storage units 45 and 46 are memories that respectively store the density data of the bright part and the density data of the dark part of the reference toner image measured when the photoconductor 11 and the like are in the initial state. The subtracter 42 calculates the difference between the density data (bright area / dark area) of the toner image (hereinafter, comparative toner image) whose density is appropriately measured and the density data stored in the initial density storage sections 45 and 46. The density change amount from the initial state is obtained, and it has a function as a density change detection means. The density change storage units 47 and 48 are memories for storing the data of the density change amount obtained by the subtractor 42.

The neurocomputer 51 as the first learning correction means is of a three-layer perceptron type, and performs calculation processing based on the data of the density change amount stored in the density change storage units 47 and 48, The exposure voltage correction amount given to the adder 35 and the charging voltage correction amount given to the adder 39 are obtained. The neuro computer 51 may be configured by a dedicated LSI or the like, or may be configured by programming a commonly used microprocessor or the like, for example.

The operation of the neuro computer 51 and the learning method will be described.

In the following description, for convenience, the neurocomputer 51 is replaced with a general three-layer perceptron type neurocomputer.

As shown in FIG. 5, the neuro computer 51 has a three-layer structure having an input layer 52, an intermediate layer 53, and an output layer 54. The input layer 52 is composed of I units, inputs values measured by the sensor, and outputs the input values as they are to the intermediate layer 53. The intermediate layer 53 is a J layer similar to the input layer 52.
Each unit is connected to each unit of the input layer 52 via a connecting portion 55 having a specific amount of connection. The units of the intermediate layer 53 are not connected to each other.

For example, as shown in FIG. 6A, the output value O2j of the j-th unit of the intermediate layer 53 is the input value of the input layer 5
2, the output value O1i of the i-th unit, the coupling amount W1ij from the i-th unit of the input layer 52 to the j-th unit of the intermediate layer 53, and the threshold value W10j of the j-th unit of the intermediate layer 53. It is expressed by the following equation using and.

[0052]

[Equation 1]

It should be noted that f is a sigmoid function shown in FIG. 7 and is a non-linear function for determining the input / output characteristics of the neurocomputer 51.

F (x) = 1 / (1 + exp (−x)) (3) The output layer 54 is composed of K units similar to the intermediate layer 53, and each unit has a unique coupling amount. It is connected to each unit of the intermediate layer 53 via the connecting portion 56. For example, as shown in (b) of FIG. 6, the output value O3k of the k-th unit of the output layer 54 is calculated from the output value O2j of the j-th unit of the intermediate layer 53 and the j-th unit of the intermediate layer 53. The amount of coupling W2jk to the k-th unit of the output layer 53 and the threshold value W20k of the k-th unit of the output layer 54 are expressed as the following equation.

[0055]

[Equation 2]

Next, the learning procedure of the neuro computer 51 configured as described above will be described with reference to the flowcharts of FIGS. 8 and 9.

Here, I and K pieces of data are given to the units of the input layer 52 and the output layer 54 as teacher data, respectively. As a result, a set of MaxGN pieces of data consisting of K pieces of data for I pieces of data is obtained.
It is assumed that there is some relationship between the I data and the K data. Further, the coupling amounts 55 and 56 of the neuro computer 51 are initially random values,
It is determined by learning on a computer from data obtained in advance by experiments.

First, the unit number i of the input layer 52 is set to 0 (S1), the unit number j of the intermediate layer 53 is set to 1 (S2), and the variable OW1 for storing the correction amount of the coupling amount 55 for one generation is set. It is cleared according to each unit number i / j (S3). OW1

[0] [j] is the intermediate layer 53
Is an area for storing the correction amount of the threshold value of each unit (hereinafter, appropriately referred to as an intermediate unit).

Next, 1 is added to j (S4), it is judged whether j exceeds the total number J of intermediate layers 53 (S5), and if j is equal to or less than J, the process returns to S3. , J
When exceeds J, 1 is added to i (S6). Further, it is judged whether or not i exceeds the total number I of input layers 52 (S7). If i is I or less, the process returns to S2. If i exceeds I in S7, OW1 Has been completed for each unit of the input layer 52 and the intermediate layer 53.

In the subsequent processing, the same as the above processing,
Variable OW for saving the correction amount of the combined amount 56 for one generation
2 is performed for each unit of the intermediate layer 53 and the output layer 54 (S8 to S14). OW2

[0] [k]
Is an area for storing the correction amount of the threshold value of each unit of the output layer 54 (hereinafter appropriately referred to as an output unit).

After that, the error storage variable r for judging the end of learning is initialized to 0 (S15), and GrpNo indicating the set number of prime definition data is initialized to 1 (S16).
When the initialization is completed, i is set to 1 (S17), and the input layer 5
The data indicated by GrpNo is input to each unit 2 (hereinafter, appropriately referred to as an input unit) (S18).

Further, 1 is added to i (S19), it is judged whether i exceeds I (S20), and if i is I or less, the process returns to S18 while i is I. When it exceeds, it means that the data input is completed for all of the input units.

When the data input is completed, the output value of each intermediate unit is calculated according to the equation (2). In this process, first, j is set to 1 (S21), and the output value O2j of the first intermediate unit is set to -W1.

[0] [j] (threshold value of the j-th intermediate unit) (S22). here,
Again, i is set to 1 (S23), and O1 [i] and W1 [i] are set.
A value obtained by adding the output value obtained as described above to the product of [j] is set as a new output value of each intermediate unit (S24).
Further, i is incremented by 1 (S25), and it is determined whether i exceeds I (S26). If i is I or less, the process returns to S24, while i exceeds I. In this case, all the output values of the first intermediate unit based on the output value from each input unit have been obtained.

Next, calculation is performed by the sigmoid function using O2j obtained in S24 as a parameter (S27).
Then, 1 is added to j (S28), it is determined whether or not j exceeds J (S29), and if j is equal to or less than J, the process returns to S22 while j exceeds J. If J
This means that the output values of all the intermediate units up to the second have been obtained.

When the output value of each intermediate unit is obtained,
In the same manner as described above, the outputs of all the Kth output units are obtained according to the equation (4) (S30 to S38). Note that W2 in the process of S31

[0] [k] is the threshold value of the kth output unit.

When the above processing is completed, the teacher data corresponding to the input data currently being calculated is put into Dt for each output unit (S39 to S42). Then, k is set to 1 (S43), and for each output unit, the squared error between the value obtained in the output layer 54 and the above-mentioned teacher data is calculated,
The result is added to r to make this a new r (S4
4). Then, the error e3 for correcting the coupling amount 56
[K] is obtained for each output unit according to the equation (5) (S45 to 47).

E3 [k] = 2 × (O3 [k] −Dt [k]) × O3 [k] × (1−O3 [k]) (5) However, in the above equation, k = 1 to K Is.

Similarly, the error e2 [j] for correcting the coupling amount 55 for each intermediate unit is obtained according to the following equation (S48-55).

[0069]

[Equation 3]

However, in the above equation, j = 1 to J. W2 [j] [k] represents the coupling amount between the jth intermediate unit and the kth output unit.

In the subsequent process, the correction amount dW2 [j] [k] of the combined amount W2 [j] [k] is calculated according to the equation (7) (S58).

DW2 [j] [k] = − ε × e3 [k] × O2 [j] (7) However, in the above equation, ε is a small value, and is usually 0.
The size is set to about 1.

When dW2 [j] [k] is obtained, the amount of binding W2 [j] [k] is corrected using this equation (8) (S59).

W2 [j] [k] = W2 [j] [k] + dW2 [j] [k] + α × OW2 [j] [k] (8) where α is a small value And is set to the same size as ε.

Further, dW2 [j] [k] at the present time is set to OW2 [j] [k] for use in the next calculation (S60). All the binding amounts 56 are corrected by the above procedure (S56 to S64), and similarly, the binding amounts W1 [i] [j] of the binding amount 55 are corrected (S65 to 73). The calculation for this correction is executed according to the equations (9) and (10).

DW1 [i] [j] = − ε × e2 [j] × O1 [i] (9) W1 [i] [j] = W1 [i] [j] + dW1 [i] [j] + α × OW1 [i] [j] (10) Then, S17 to S for the MaxGN teacher data.
If the square error sum r is less than R, the neurocomputer 51 is considered to have learned the relationship between the input data group and the output data group of the teacher data (S76), and the processing is performed. Ends. If r is R or more, the processing from S15 is repeated.

In the neurocomputer 51 which has completed learning as described above, a flowchart showing the processing procedure for outputting an expected value for given data is shown in FIG.
This will be described with reference to the flowchart of

First, input data is set to I input units (S101 to S104), and the output value of each intermediate unit is calculated according to the equation (2) (S105 to S11).
3). Next, the output value of each output unit is calculated according to the equation (4) (S114 to S122). The above calculation is the same as the processing of S18 to S38 in the learning procedure described above.

After the learning is completed, the neurocomputer 51 determines the relationship between the input / output data of the teacher data and the coupling amount 55 thereof.
-It is acquired in the form of 56, and outputs a value expected from the input / output relationship peculiar to the teacher data with respect to an appropriate input value. This function operates almost correctly even for input values that have not been learned, and outputs an almost accurate expected value for any input value if the input value range is appropriate.

On the other hand, in the neurocomputer 51 shown in FIG. 1 which is applied to this copying machine, the input layer 52 is composed of two units 52a and 52b.
2b, the density change storage unit 4 is used as input data.
The output value of 7.48 is input. In addition, the neuro computer 51 includes units 54a and 5 having two output layers 54.
4b, the unit 54a outputs the exposure voltage correction amount, and the unit 54b learns to output the charging voltage correction amount.

In learning the neuro computer 51, first, several kinds of conditions such as temperature / humidity and the number of times of use are selected, and under these conditions, two kinds of densities of the reference original 21 are used to set the reference toner on the photoconductor 11. An image is formed, and the toner density sensor 22 measures the amount of change in toner density from the initial state. At this time, the exposure voltage and the charging voltage are set as preset fixed voltages to form an image. At the same time, the correction amounts of the exposure voltage and the charging voltage required to obtain the optimum copy image quality are obtained while actually making the copy. By using these correction amounts as teacher data and performing learning on a computer according to the procedure shown in FIGS. 8 and 9, the final combined amounts 55 and 56 are obtained and the learning is completed.

In the copying machine equipped with the neuro computer 51, first, the toner density is measured once when the photoconductor 11 and the like are in the initial state, and the density is stored in the initial density storage unit 45.
46. After that, the toner density is appropriately measured, and the density change amount from the toner density in the initial state is stored in the density change storage units 47 and 48. The neurocomputer 51 outputs an expected value according to the procedure shown in FIG. 10. However, in the present embodiment, as described above, the output value of the unit 54a of the output layer 54 becomes the exposure voltage correction amount and the unit 54b. Is learned so that the output value of is the charging voltage correction amount.

Therefore, the voltage stored in the exposure voltage storage section 36 is corrected by the exposure voltage correction amount, and the exposure amount is appropriately controlled. In addition, the charging voltage storage unit 4
The voltage stored in 0 is corrected by the above charging voltage correction amount, and the charging potential of the photoconductor 11 is appropriately controlled. By performing such control, "○" is shown in FIG.
As shown in, the simulated original density vs. toner density characteristics are very close to the initial characteristics shown by the solid line.

As described above, according to the present embodiment, the exposure characteristics and the charging characteristics can be optimally controlled by the neuro computer 51, and the copy image quality can always be kept good. Further, in this embodiment, since the correction is performed based on the measured density of the reference toner image, the only sensor required for that purpose is the toner density sensor 22, and
The number of sensors can be reduced as compared with a control method in which correction is performed while measuring the environment such as humidity. Further, in the present embodiment, the exposure voltage and the charging voltage are corrected by learning based on the reference toner density, so that the exposure amount and the change in the charging voltage of the photoconductor 11 due to the contamination of the exposure lamp 8 and the charger 12 can also be corrected. Therefore, the optimum image quality control can be performed.

Although the exposure voltage and the charging voltage are used as the control values in the present embodiment, the present invention is not limited to this, and the developing bias voltage may be used as the control value. The combination of these control values is selected according to the performance of each copying machine, and is not limited to the combination of this embodiment.

[Embodiment 2] Another embodiment in which the present invention is applied to a copying machine will be described below with reference to FIGS. 2 and 12. The constituent elements in this embodiment that have the same functions as those of the constituent elements described in the first embodiment will be designated by the same reference numerals and the description thereof will be omitted.

As shown in FIG. 2, the copying machine according to this embodiment has the same basic structure as the copying machine of the first embodiment. Further, this copying machine is designed to control the image quality by correcting the outputs of the exposure lamp 8 and the charger 12 as in the copying machine of the first embodiment. The cumulative operating status of the electrophotographic process in the machine is added.

As shown in FIG. 12, the control system of this copying machine includes a developing counter 61 connected to the developing device 13. Development counter 61 as operating status detecting means
Is a counter that adds 1 to the count value each time the electrophotographic process is performed once and the development is performed once, and the count value is used as the cumulative usage status. The developing counter 61 is composed of a non-volatile memory or the like, and when the developer in the developing device 13 is replaced with a new one,
Cleared to 0 by maintenance personnel.

Further, the present copying machine is provided with a neuro computer 62 in the control system for controlling the image quality. Second
The neuro computer 62 as the learning correction means has three
An input layer 63 including a plurality of units 63a to 63c, an intermediate layer 64 including a plurality of units, and two units 6
An output layer 65 composed of 5a and 65b is provided. The units 63a and 63b respectively include a density change storage unit 47
48 is connected, and the development counter 6 is connected to the unit 63c.
1 is connected. Further, adders 35 and 39 are connected to the units 65a and 65b, respectively.

The neuro computer 62 is the neuro computer 51 (see FIG. 1) in the copying machine of the first embodiment.
Similarly to the above, the exposure voltage correction amount and the charging voltage correction amount are corrected based on the output values of the density change storage units 47 and 48, but the above count value is taken in as an element for the correction. It is like this.

When the neuro computer 62 is learned, the reference toner image is created and the density change amount between the reference toner image and the comparative toner image is obtained in the same manner as the neuro computer 51. Then, by using the result and the count value of the developing counter 61 as the teacher data, learning is performed, and the final combined amount 66/67 is obtained, and the learning is completed.

In this copying machine, since the count value of the developing counter 61 is used as an element for correcting the exposure voltage and the charging voltage as described above, deterioration of the developer may be indirectly predicted. As a result, the correction can be performed with higher accuracy.

In this embodiment, the cumulative usage status is detected by counting the number of times of development.
The development counter 61 may be replaced as long as the number of developments can be counted by the number of copying operations. For example, a configuration is possible in which a lighting counter that counts the number of times the exposure lamp 8 is lit is provided and the count value is used. The lighting counter, like the developing counter 61, is cleared to 0 by maintenance personnel when the developer is replaced with a new one. Further, the number of inputs of the neuro computer 62 may be four, and the output values of the developing counter 61 and the above lighting counter may be input in parallel.

If the toner density measurement in this embodiment is carried out, for example, after the power of the copying machine is turned on until the temperature of the fixing device 15 reaches a prescribed temperature, the original copying operation will be performed. Since it does not interfere with, there is almost no practical inconvenience.

[0095]

As described above, the apparatus for stabilizing the image quality of the electrophotographic apparatus according to the first aspect of the present invention includes the density detecting means for detecting the density of the toner image formed on the photoconductor and the density detecting means. And a density change detecting means for detecting a change in the density of the toner image, the density of which has been changed by the execution of the electrophotographic process, with respect to the reference density of the toner image in which the density is in the initial state. When the density change is input, the correction amount of the control value is input based on the density change detected by the control unit and the correction amount of the control value that affects the image quality in the electrophotographic process. In addition to learning the processing procedure for outputting, the predicted change amount of the control value according to the density change is predicted according to the learned processing procedure by the density change newly input after learning. Is a configuration that includes a first learning correction means for force.

As a result, since the control value correction amount is obtained by the processing procedure based on learning, it is possible to collectively correct a plurality of control values. Therefore, the correct control value can be corrected even on the basis of the density change including the density change due to dirt on the exposure lamp or the like. Further, since only the density detecting means for detecting the density of the toner image is newly provided in the electrophotographic process section, it is not necessary to increase the number of sensors and accordingly to secure a space for the sensor in the electrophotographic process section.

Therefore, if the image quality correction device for an electrophotographic apparatus according to claim 1 is adopted, an image of stable image quality can always be obtained by the above correction, and an image is newly provided in the electrophotographic process section. It is possible to reduce the number of parts such as a sensor and suppress an increase in cost of the electrophotographic apparatus.

As described above, the image quality stabilizing device for the electrophotographic apparatus according to the second aspect of the present invention detects the density of the toner image formed on the photoconductor by the density detecting means and the density detecting means. The density change detection means for detecting a change in the density of the toner image whose density has changed due to the execution of the electrophotographic process with respect to the reference density of the toner image in the initial state, and the toner image density in the initial state. Operating status detecting means for detecting the cumulative operating status of the electrophotographic process from a certain time, and the cumulative operating status of the electrophotographic process detected by the density change and the operating status detecting means detected by the density change detecting means, Based on the density change of the control value affecting the image quality in the electrophotographic process and the correction amount according to the cumulative operating condition, the density change and When the cumulative operating status is input, the processing procedure for outputting the correction amount of the control value corresponding thereto is learned in advance, and the learned processing is performed by the newly input density change after the learning and the cumulative operating status of the electrophotographic process. Second learning correction means for predicting and outputting the correction value of the control value according to the density change and the cumulative operation status of the electrophotographic process according to the procedure.

As a result, like the image quality correction device for an electrophotographic apparatus according to claim 1, a plurality of control values can be corrected in an integrated manner, and it is not necessary to increase the number of sensors or secure a space accordingly. . Further, in the above configuration, since the control value is corrected even based on the cumulative operation status of the electrophotographic process as described above,
It is also possible to predict a change in image quality due to deterioration of the developer.

Therefore, if the image quality correction apparatus for an electrophotographic apparatus according to claim 2 is adopted, in addition to the effect of the image quality correction apparatus for an electrophotographic apparatus according to claim 1, the image quality is corrected with higher accuracy. There is an effect that can be.

[Brief description of drawings]

FIG. 1 is a block diagram showing a control system of a copying machine according to a first embodiment of the present invention.

FIG. 2 is a front view showing the internal structure of the copying machine according to the first and second embodiments of the present invention.

FIG. 3 is a perspective view showing a reference document provided in the copying machine of FIG.

FIG. 4 is a circuit diagram showing a configuration of a toner concentration sensor provided in the copying machine of FIG.

5 is an explanatory diagram showing a configuration of a neuro computer in the control system of FIG.

6 is an explanatory diagram showing a configuration of units in an intermediate layer and an output layer of the neurocomputer shown in FIG.

7 is a graph showing a sigmoid function used for calculating respective output values of an intermediate layer and an output layer in the neurocomputer of FIG.

8 is a flowchart showing a processing procedure at the time of learning of the neurocomputer of FIG.

9 is a flowchart following the flowchart of FIG. 8 showing a processing procedure at the time of learning of the neurocomputer of FIG.

10 is a flowchart showing a processing procedure at the time of image quality correction of the neurocomputer of FIG.

FIG. 11 is a graph showing the characteristics of toner density with respect to original density, showing the initial characteristics and the characteristics obtained by simulation assuming the copying machine according to the first embodiment of the present invention and the conventional image stabilizing apparatus. It is a graph which compared with the characteristic acquired by simulation assuming.

FIG. 12 is a block diagram showing a control system of a copying machine according to a second embodiment of the present invention.

FIG. 13 is a graph showing document density-exposure light amount characteristics, charging potential-exposure light amount characteristics, charging potential-toner density characteristics, and document density-toner density characteristics required for image quality correction in a conventional image stabilizing device.

[Explanation of symbols]

8 Exposure Lamp 11 Photosensitive Member 12 Charging Device 13 Developing Device 21 Reference Document 22 Toner Density Sensor (Density Detecting Means) 42 Subtractor (Density Change Detecting Means) 45/46 Initial Density Storage 47/48 Density Change Storage 51 Neurocomputer (First learning correction means) 61 Development counter (operating status detection means) 62 Neurocomputer (second learning correction means)

Claims (2)

[Claims] 1. A density detecting means for detecting the density of a toner image formed on a photoconductor, and an electrophotographic process for a reference density of a toner image whose density is in an initial state detected by the density detecting means. Density change detecting means for detecting a change in the density of the toner image whose density has changed by the density change detected by the density change detecting means, and the density change of the control value affecting the image quality in the electrophotographic process. On the basis of the correction amount according to, the processing procedure for outputting the correction amount of the control value according to the input of the density change is learned in advance, and the learning is performed by the density change newly input after learning. And a first learning correction means for predicting and outputting the correction amount of the control value according to the density change according to the processing procedure described above. Quality stabilization device. 2. A density detecting means for detecting the density of a toner image formed on a photoconductor, and an electrophotographic process for a reference density of a toner image whose density is in an initial state detected by the density detecting means. A density change detecting means for detecting a change in the density of the toner image in which the density has changed, and an operating status detecting means for detecting the cumulative operating status of the electrophotographic process since the toner image density is in the initial state, The density change detected by the density change detecting means and the cumulative operating status of the electrophotographic process detected by the operating status detecting means, the density change of the control value affecting the image quality in the electrophotographic process and the cumulative operating status When the density change and the cumulative operating condition are input, the correction amount of the control value is output according to the correction amount. In addition to learning the processing procedure in advance, the density change and the cumulative operating status of the electrophotographic process newly input after learning allow the control of the control value according to the density change and the cumulative operating status of the electrophotographic process according to the learned processing procedure. Second to predict and output the correction value
An image quality stabilizing device for an electrophotographic apparatus, comprising: learning correction means.