JPH06253151A - Image sampling correction system for register matching for multiplexed image forming device - Google Patents

Abstract

PURPOSE:To attain high speed sampling and smoothing arithmetic operation processing by improving a register matching image sampling correction system for the multiplexed image forming device. CONSTITUTION:The system is provided with each image output means 8, its output control means 7, a pattern generator 6 generating repetitively a rough or an accurate pattern for register deviation measurement in the main scanning direction and subscanning direction respectively, a sampling means 3 and a sampling control means 2 executed at both sides of a transfer belt 9 in the main scanning direction, an arithmetic operation processing means 5 receiving sampling data and calculating register deviation, and a sampling data storage means 4, and when image sampling correction is controlled, the control means 1 selects a register deviation measurement pattern generated from the pattern generator in response to the mode as to whether it is the rough adjustment mode or the fine-adjustment mode, sets a sample start point and a sample width of the sampling control means 2, generates the repetitive register deviation measurement pattern and integrates sampling data or arithmetic operation processing data to obtain a pattern position.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a registration image sampling / correction method for registering a multiple image forming apparatus which forms a multiple image by arranging a plurality of image output means along a transfer belt. .

[0002]

2. Description of the Related Art In recent years, colorization of copiers, printers, fax machines, etc., which handle documents has progressed rapidly, and documents in offices have been colorized. Devices that handle these color documents tend to be further speeded up in the future.

As a device for handling color documents,
For example, four ROSs (Raster Outpu) for each color of black (K), yellow (Y), magenta (M), and cyan (C).
A so-called tandem color printer having a tscanner has been proposed (for example, JP-A-1-142674, JP-A-1-142680, JP-A-1-18).
3676, JP-A-1-281468, etc.).
However, since this tandem color printer is a method in which a plurality of separate ROSs form one image, a paper jam or other abnormal operation may cause a user or a service person to make a part of the image forming apparatus. Temporarily move from the original image forming position, return to the original position after exchanging those parts, and if there is a change in temperature, a change with time, or an impact, these two positions There is a big problem that a subtle error occurs in the relationship and a misregistration of each color occurs.

The cause of the registration deviation is a deviation component in the main scanning direction of scanning by the ROS, a deviation component in the conveying direction of the transfer member conveying belt, that is, a deviation component in the sub-scanning direction, and an expansion / contraction of an image in the scanning direction of the ROS, ie, ROS There are scan magnification shift components, ROS scan angle shifts, that is, ROS skew shift components.

Therefore, the image position recognition pattern predetermined by each ROS is output from the pattern generator according to a certain rule, and the image position recognition pattern is predetermined by the CCD arranged downstream of those ROSs. Sampling at timing. Actually CC
The positional relationship expected to be obtained when it is assumed that there is no color shift of the image position recognition pattern of each color determined in advance in the positional relationship of the sampling data of the image position recognition pattern of each color sampled by D, The amount of difference is detected, and the registration deviation amount of each color is calculated from the detected data. A method has been proposed in which a high quality image with less registration deviation is provided by correcting the write timing of the ROS based on the result.

[0006]

However, the conventional method has the following problems when the registration deviation measuring pattern output from each ROS according to a certain rule is sampled by the CCD arranged downstream thereof. was there.

[0007]. In order to increase the accuracy of detecting the registration shift amount, the sampling speed of the CCD must be increased. Further, in order to perform highly accurate registration shift correction in finer steps, it is necessary to further increase the sampling speed of the CCD. However, if the sampling speed of the CCD is increased, the CPU peripheral memory, which usually has a slow access speed, cannot be used. So CP
If a high-speed memory is used as the U peripheral memory, the CPU cannot use the bus.

[0008] In a registration deviation measuring system that measures both the registration deviation measuring pattern in the main scanning direction and the registration deviation measuring pattern in the sub-scanning direction, for example, during sampling of some registration deviation measuring patterns in the sub-scanning direction. When sampling of the registration deviation measurement pattern in the main scanning direction occurs, the image data processing of the registration deviation measurement pattern in the main scanning direction is performed on top of the image data processing of the registration deviation measurement pattern in the sub-scanning direction. Will be added to the form. Therefore, the calculation processing of the CPU is concentrated,
It is necessary to use an expensive CPU or to keep the operating rate of a normal CPU low. Further, since the sampling interval (pattern cycle) of the registration deviation measurement pattern is limited where the calculation processing of the CPU is concentrated, there is a drawback that it is difficult to capture a color registration vibration component having a high frequency. It was

[0009]. When a plurality of CCDs are used to sample the registration deviation measurement pattern, if one RAM is shared and used, it is necessary to strictly manage so that the timings at which the CCDs register the registration deviation measurement pattern do not overlap. As a result, the pattern interval cannot be shortened for the purpose of capturing the color registration vibration component having a high frequency and fluctuating.

[0010]. When an image position is determined after sampling the registration displacement measurement pattern, for example, if a sophisticated image position determining method using the center of gravity method or the like is adopted, the time required for the calculation and processing becomes enormous. As a result, an enormous amount of memory is required, the time required for the calculation and processing after sampling is enormous, and the processing time is too long to catch up with. There was a drawback that it was difficult to capture the ingredients.

Furthermore, in order to reliably sample the registration deviation measuring pattern in which a fixed positional deviation (hereinafter referred to as a DC component deviation) and a variable positional deviation (hereinafter referred to as an AC component deviation) peculiar to the apparatus have occurred, an output from the CCD is output. Sampling the data increases the sample width, requires a huge amount of memory, and requires a huge amount of data processing time after sampling to process the data.

Further, in the tandem color printer, the AC component deviation, such as the deviation of the conveyor belt speed, the deviation of the conveyor belt walk (deviation in the main scanning direction), and the deviation of the transfer drum rotation, are so large that they cannot be ignored. If these AC component deviations are not taken, the original purpose DC component deviation cannot be measured. A method of obtaining the AC component shift is, for example, a register shift measurement pattern having a width twice as large as a frequency (known sampling theorem) that is not easily influenced by the shift frequency, and is repeated to obtain an average to obtain a DC
Only the component deviation can be measured. However, in the conventional method, since the sample width for reliably sampling the registration deviation measurement pattern is large, the pattern pitch of each of the registration deviation measurement patterns also becomes large,
The width of the registration shift measuring pattern itself could not be reduced.

The present invention has been made to solve the above-mentioned problems, and improves the detection accuracy of the image position address and the registration deviation amount of the registration deviation measurement pattern, and makes it possible to perform highly accurate registration deviation correction. Is to provide a registration registration image sampling correction method. Another object of the present invention is to reduce the amount of sample data by efficiently controlling the sample timing so that the image position address can be accurately detected. Another object of the present invention is to enable high-speed sampling and smooth operation processing of the CPU.

[0014]

To this end, the present invention provides a registration image for registration of a multiple image forming apparatus that forms a multiple image by arranging a plurality of image output means along a transfer belt. In the sampling correction method, output control means for supplying image data to each image output means to control image formation, and coarse registration deviation measurement pattern or fine registration deviation measurement in the main scanning direction and sub-scanning direction of each image output means. Generator for repeatedly generating the use pattern, sampling means for sampling the image formed by each image output means on both sides of the transfer belt in the main scanning direction, and sampling control for controlling the sample start point and sample width of the sampling means. Means and the data sampled by the sampling means are taken in Output processing, data storage means for storing sampling data or arithmetic processing data, and control means for controlling image output and image sampling correction. When controlling the sampling correction, the registration deviation measurement pattern generated by the pattern generator is selected according to the coarse adjustment mode or the fine adjustment mode, and the sampling start point and the sample width of the sampling control means are set to repeat the registration adjustment. The present invention is characterized in that a deviation measuring pattern is generated and sampling data or arithmetic processing data is integrated to obtain a pattern position.

Further, as the data storage means, there is provided a dedicated memory for storing the image data of the registration deviation measuring pattern in the main scanning direction and a dedicated memory for storing the image data of the registration deviation measuring pattern in the sub-scanning direction. And

Further, the arithmetic processing means preferentially processes the image data of the shorter sampling period, divides the remaining image data to distribute the processing, and sets the reference position of the registration deviation measuring pattern to each sample. Image data is integrated with the deviation from the reference position by correcting each start point, or it is discriminated whether or not the image data is a registration displacement measurement pattern, and the processed data and exclusion data are identified by the discrimination result of the registration displacement measurement pattern. Is characterized in that

The control means corrects the sample start point on the basis of the image position information of the previous pattern, makes the interval between the sample start points of the registration deviation measurement pattern constant, and makes the image of the previous pattern for each cycle. The sampling start point is corrected on the basis of the position information, a coarse registration deviation measuring pattern is selected in the coarse adjustment mode, and control is performed so as to capture sampling data during one rotation of the photosensitive drum of the image output means. Alternatively, in the fine adjustment mode, a fine registration shift measuring pattern is selected, and control is performed so as to capture sampling data during one rotation of the transfer belt.

[0018]

In the registration registration image sampling correction method of the present invention, a pattern generator that repeatedly generates a coarse registration deviation measurement pattern or a fine registration deviation measurement pattern in the main scanning direction and the sub scanning direction of each image output means is used. When controlling the image sampling correction, the control unit selects a registration deviation measurement pattern generated by the pattern generator according to the coarse adjustment mode or the fine adjustment mode, and also sets the sample start point and the sample width of the sampling control unit. The pattern start position and the sample width are set properly even if the number of samplings is increased, because the pattern position is calculated by accumulating the sampling data or the calculation processing data in the data storage means by generating the pattern for repeated registration misalignment measurement and storing it in the data storage means. Can be done in the cash register It is possible to improve the correction accuracy.

Further, as the data storage means, a dedicated memory for storing the image data of the registration deviation measuring pattern in the main scanning direction and a dedicated memory for storing the image data of the registration deviation measuring pattern in the sub-scanning direction are provided. Sampling becomes possible.

Further, the arithmetic processing means preferentially processes the image data of the shorter sampling period, divides the remaining image data to distribute the processing, and sets the reference position of the registration deviation measuring pattern to each sample. Image data is integrated with the deviation from the reference position by correcting each start point, or it is discriminated whether or not the image data is a registration displacement measurement pattern, and the processed data and exclusion data are identified by the discrimination result of the registration displacement measurement pattern. Therefore, it is possible to smooth the arithmetic processing time, reduce the amount of processing per unit time, and eliminate useless sample data.

The control means corrects the sample start point on the basis of the image position information of the previous pattern, makes the interval between the sample start points of the registration deviation measurement pattern constant, and makes the image of the previous pattern for each cycle. Correcting the sample start point based on the position information, selecting a coarse registration deviation measurement pattern in the coarse adjustment mode, and controlling so as to capture sampling data during one rotation of the transfer drum of the image output means. , Or by selecting a fine registration displacement measurement pattern in the fine adjustment mode and controlling so as to capture the sampling data while the transfer belt makes one revolution, it is possible to finely adjust the pattern so that the pattern can be made finer. Furthermore, the spread can be made smaller and high AC components can be detected without increasing the amount of calculation.
Sample efficiency can be improved without losing sight of the pattern within the spread range.

[0022]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a configuration of an embodiment of a registration registration image sampling correction method of a multiple image forming apparatus according to the present invention, FIG. 2 is a diagram showing an example of a coarse adjustment pattern, and FIG. 3 is a diagram showing an example of a fine adjustment pattern. 4A and 4B are views for explaining how to correct the registration deviation correction factor, and FIG. 5 is a view for assembling and SKE.
FIG. 6 is a diagram for explaining a registration adjustment algorithm at the time of W coarse correction, FIG. 6 is a diagram for explaining a coarse adjustment cycle algorithm, FIG. 7 is a diagram for explaining a fine adjustment cycle algorithm, and FIG. 8 is a registration check cycle. 3 is a diagram for explaining the algorithm of FIG.

In FIG. 1, the transfer belt 9 is a belt-shaped transfer member for transferring the image formed by the image output section 8. The image output section 8 has a plurality of image output means along the transfer belt 9. Are arranged to form a multiple image. The image control unit 7 supplies image data to the image output unit 8 to control the image formation. The image output unit 8 has rough adjustment patterns in the main scanning direction and the sub scanning direction as shown in FIG. This is for repeatedly generating a fine registration shift measurement pattern as shown in FIG. 3 as a registration shift measurement pattern or a fine adjustment pattern. The sampling unit 3 uses the image output unit 8 to transfer the transfer belt 9
The sampling control unit 2 controls the sampling start point and the sampling width of the sampling unit 3 by, for example, a CCD sensor that samples the image formed on the both sides in the main scanning direction. Arithmetic processing unit 5
Is for taking in the image data sampled by the sampling section 3 and performing arithmetic processing for calculating the registration deviation, and the data storage section 4 is for storing sampling data or arithmetic processing data. The control unit 1 controls image output and image sampling correction. When controlling image sampling correction, the control unit 1 measures the registration deviation generated by the pattern generator 6 according to the coarse adjustment mode or the fine adjustment mode. In addition to selecting a pattern, a sampling start point and a sample width of the sampling controller 2 are set to repeatedly generate a registration deviation measurement pattern, and sampling data or arithmetic processing data is integrated to obtain a pattern position. Then, the misregistration is detected, and the image control unit 7 and the image output unit 8 are controlled to perform skew correction, magnification correction, axial direction correction, density correction, and the like correction.

The misregistration measurement pattern includes a rough pattern for the initialization cycle shown in FIG. 2 for coarse adjustment and a fine pattern for the registration correction cycle shown in FIG. 3 for fine adjustment, each of which is constant in the main scanning direction. And an image having a fixed length in the sub-scanning direction. In the case of the rough adjustment pattern, for example, as shown in FIG. 2, a line having a raster width of a raster width on one side from the center of the CCD sensor of the sampling unit 1 and a length of bmm in the main scanning direction and the opposite side. Is developed so that K, Y, M, and C are connected in series at d raster intervals as a set with a line having a length of c raster in the sub-scanning direction. On the other hand, in the case of the pattern for fine adjustment, for example, as shown in FIG.
Lines having a length of k raster lines are arranged in the sub-scanning direction in the order of K, Y, M, and C at f raster intervals, and a line having a length of g raster lines in the sub-scanning direction is formed on the side thereof (on the opposite side). Each time the set is repeated, lines in the sub-scanning direction are developed in the order of K, Y, M, and C.

In the registration shift correction, for example, when there is a shift like the image Y with respect to the image K as shown in FIG. 4A, the mirror angle other than K is first corrected with reference to the image K. Skew adjustment to adjust the angles as shown in (b), and then the magnification is corrected by controlling the clock frequency to adjust the main scanning direction magnification as shown in (c) to adjust the write clock timing and phase. The main scanning direction is corrected by (1) to match the main scanning direction, and the line sync writing timing and polygon phase correction are used to correct the sub-scanning direction as shown in (e), and finally the allowable error Correct within.

Therefore, in the registration adjustment at the time of assembly and SKEW coarse correction, as shown in FIG. 5, a coarse adjustment cycle is first executed (step S11), and SKEW coarse correction and coarse correction data other than SKEW are transmitted to perform registration. Displacement correction is executed (step S12).

At this time, the SKEW is corrected by the SKEW coarse correction.
Since deviation occurs in elements other than W, the rough adjustment cycle is repeatedly executed again (step S13), and coarse correction data other than SKEW is transmitted to execute registration deviation correction (step S14).

Next, a fine adjustment cycle is executed (step S
15), SKEW fine correction and fine correction data other than SKEW are transmitted to execute registration deviation correction (step S16).
The SKEW correction ends for the first time here, but this SKEW correction
Even with fine correction, the elements other than SKEW are misaligned. Therefore, for elements other than SKEW, the coarse adjustment cycle is executed again (step S17), and the coarse adjustment data other than SKEW is transmitted (step S18). The cycle is executed (step S19), fine correction data other than SKEW is transmitted, and registration deviation correction is executed (step S2).
0).

In the coarse adjustment cycle of the above processing, sampling data of the coarse adjustment pattern is fetched as shown in FIG. 6 (steps S21 to S22), and the sampling data is calculated to obtain the image position (step S23). ,
When the image position for all sampling data is obtained (step S24), the correction data is calculated (step S24).
25), and the correction data is transmitted (step S26).

In the fine adjustment cycle, as shown in FIG. 7, similarly to the coarse adjustment cycle, the sampling data of the fine adjustment pattern is taken in (steps S31 to S32), and the sampling data is calculated to obtain the image position (step S3).
3) When the image positions of all the sampling data are obtained (step S34), the correction data is calculated (step S35) and the correction data is transmitted (step S3).
6).

Then, in the checkout cycle, FIG.
When the registration check pattern is sampled (step S41), the sampling is completed (step S42), the image position is calculated (step S43), and the image position calculation is completed for all samplings. (Step S44), the registration deviation is calculated (step S45), and it is checked whether or not the registration deviation amount is within the range of the specified value A (step S46). It is checked whether or not it is within the range of the large specified value B (step S47). If it is within the range, the fine adjustment cycle is executed, and if it is out of the range, the coarse adjustment cycle and the fine adjustment cycle are continuously executed (steps S48-).
S50).

Further, the device configuration of the present invention will be described in detail. FIG. 9 is a diagram showing a configuration example of a registration registration image sampling correction device, and FIG. 10 is a block diagram showing a configuration example of a registration deviation correction substrate.

In FIG. 9, a CCD sensor 11 reads an image formed on the transfer belt 7 by the image forming devices 13K, 13Y, 13M and 13C, one at each end of the image area, for a total of two. Will be placed. Light source 12
Is a light source for producing the background light necessary for the CCD sensor 11 to detect the image on the transfer belt 17, and L
Any element such as an ED or a halogen lamp can be used as long as it can secure a sufficient amount of light as the light source of the CCD sensor 11. In addition, the light source 12 is deteriorated in its light amount, and the transfer belt 17
It is possible to freely change the amount of light in order to secure the optimum image receiving state against the deterioration of the transmittance of the CCD, the deterioration of the sensitivity of the CCD sensor, the deterioration of the transmittance due to dirt on the optical system, and the environmental change represented by the temperature. is there. Image forming apparatus 13K, 13
Y, 13M, and 13C are, for example, a combination of a photoconductor and a laser beam ROS or LED ROS, which are arranged at substantially equal intervals to form black images,
A device for forming a yellow image, a device for forming a magenta image, and a device for forming a cyan image.

The transfer belt 17 includes an image forming device 13K,
It has a transparent belt structure for transferring an image formed by 13Y, 13M and 13C, and is supported by a driving roller 19a and a driven roller 19b facing the driving roller 19a. Further, the transfer belt 17 has a function of carrying the transfer paper, and the transfer paper is carried from right to left in the drawing in the same manner as the belt rotating direction. At this time, a suction corotron (not shown) is provided to suck the transfer sheet onto the belt member. The drive roller 19a is driven by a dedicated drive motor (not shown) having excellent constant speed,
The driven roller 19b rotates as its driving force is transmitted by the transfer belt 17. The assay body 20 is C
The CD sensor 11 and its drive circuit, an optical system such as a SELFOC lens, etc. are mounted, and the mounting relationship between them is designed so that highly accurate positioning can be easily realized.

Interface boards 14K, 14Y, 14
M and 14C are for sending an image signal to the ROS. The registration deviation correction board 15 is a board for accommodating the module of the registration deviation correction system, and the control board 16 is for managing the operation of each board and the entire apparatus. The board for housing the module and the image processing board 21 is a board for housing the memory and the module related to the image processing.

Next, the normal image forming mode will be described. When the leading edge of the sheet conveyed by the transfer belt 17 reaches the transfer point directly below the image forming apparatus 13K,
A state in which the leading edge of the image formed by the image forming apparatus 13K reaches a transfer point directly below the image forming apparatus 13K, that is, the sub-scanning direction between the image formed by the image forming apparatus 13K and the sheet (sheet conveying direction). The paper feed timing and the image writing timing are determined so that there is no misalignment. Although not shown, the image formed by the image forming apparatus 13K is transferred to the sheet that has reached the copy point by a transfer corotron, and further reaches the transfer point directly below the image forming apparatus 13Y. The sheet that has reached the transfer point directly below the image forming apparatus 13Y is transferred in the same manner as it was transferred by the image forming apparatus 13K. Hereinafter, the image forming apparatus 1
Similarly, 3M and 13C are also transferred in a superimposed manner. The sheet after all the transfer is further conveyed by a belt, and when it reaches the vicinity of the driven roller 19b, it is separated from the transfer belt 17 by a corotron, a stripper or the like for separating the sheet from the transfer belt 17, although not shown. After that, it is fixed by a fixing device or the like, and discharged outside the machine.

Next, the registration deviation correction system for each color will be described. The registration shift correction is executed by entering a dedicated correction cycle preset in the apparatus. When a command is issued from the control board 16 to each board, this correction cycle starts. When the registration deviation correction board 15 prepares for sampling the registration deviation measurement pattern and the correction cycle starts, the interface boards 14K and 14K
Y, 14M, and 14C play the role of a pattern generator that outputs a registration deviation measurement pattern, and the registration deviation measurement pattern is output to each of the interface boards 14K, 14Y, and
14M, 14C to the image forming apparatuses 13K, 13Y, 13
M, 13C, pattern 18K, 18Y, 18
M and 18C are transferred onto the transfer belt 17.

There are two types of registration deviation measuring patterns for rough adjustment (for initialization cycle) as shown in FIG. 2 and for fine adjustment (for registration correction cycle) as shown in FIG. The registration deviation measurement pattern is set so as to be transferred to a position on the transfer belt 17 that passes directly under the CCD sensor 11. Then, in the registration deviation correction board 15 for sampling the image data from the CCD sensor 11, the interface boards 14K, 14Y,
Transfer timing and CCD sensor 11 at 14M and 14C
The sampling start timing and the sampling end timing which are necessary and sufficient for sampling the registration deviation measuring pattern are determined from the pitches up to and the image data is captured.

In the registration error correction substrate 15, first, when the sampling start timing comes, the image data from the CCD sensor 11 starts to be taken into the high-speed memory, and when the sampling end timing comes, the reading of the image data is stopped. Then, at the same time as the end of the acquisition, by the time the sampling of the next registration deviation measuring pattern is completed, the image position is determined from the acquired image data by, for example, the center of gravity method, and used as the image position address, for example. Store in main memory. By repeating this operation several times, several fixed image position addresses are obtained for each image forming apparatus. Here, in order to improve the accuracy of the fixed image position address, these several fixed image position addresses may be averaged for each image forming apparatus.

Further, in the registration deviation correction substrate 15, a correction value for correcting the registration deviation between the image forming apparatuses is corrected by a predetermined algorithm based on the image position address determined for each image forming apparatus. It is calculated for each image forming apparatus. The calculated correction values are directly or indirectly set on the image forming apparatus, the interface board, or the like from the registration deviation correction board 15 to perform the correction.

The sampling start timing necessary and sufficient for sampling the registration deviation measurement pattern is based on the conveyor belt speed and the expected deviation range from the timing when the registration deviation measurement pattern is output by the image forming apparatus. It is determined by the time required to reach directly under the CCD sensor 11. If the sampling interval is lengthened in order to always sample the registration deviation measurement pattern well, the storage area for storing the sampling data becomes large. Therefore, first, a registration deviation measuring pattern having a wide interval is used as shown in FIG. 2, for example, rough adjustment is performed as an initialization cycle in one round of the drum, and then a registration deviation measuring pattern having a small interval is measured as shown in FIG. Using a pattern, fine adjustment is performed as a registration correction cycle for one round of the belt. Neither fine adjustment nor coarse adjustment can correct the AC component. However, at the time of rough adjustment, only the DC component including the AC component error can be corrected because the sample amount is small. In fine adjustment, a large amount of samples are made at high frequencies, so that DC components can be corrected by averaging them. Therefore, in both the coarse adjustment and the fine adjustment, the deviation of the DC component is made to approach 0.

Further, for example, when correcting the registration deviation of each of the other ROS outputs with reference to the first ROS (K) output,
The first ROS may not have a correction function. In that case, even if the rough sampling is performed and the correction is performed, the sampling may not be performed in the next fine sampling. In such a case, after the coarse sampling, the deviation of the output of the first ROS is monitored, and the registration deviation measurement pattern is measured by the CCD.
Correcting the time to reach directly under the sensor allows fine sampling without error.

The registration shift correction substrate 15 is constructed, for example, as shown in FIG. In the registration shift correction substrate 15, C
The driver 32 drives the CCD sensor according to the clock generated by the CD drive clock generation circuit 38,
For example, read image data of 8 bits and 256 gradations is sequentially taken into the receiver 31 in pixel units. Then, the image data regarding the main scanning is stored in the high-speed image memory for main scanning 35 through the bus control system 34, and the image data regarding the sub-scanning is averaged by the image arithmetic circuit for sub-scanning 33, and then the bus control system. It is stored in the sub-scanning high-speed image memory 36 through 34. Sample timing control circuit 39
Is the sampling start timing set by the CPU 44,
The sub-scanning image calculation circuit 33 according to the sample period or the like.
Also, the timing for loading image data into the main scanning high-speed image memory 35 and the sub-scanning high-speed image memory 36 is controlled. The main RAM 42 is used as a work area for the CPU 44, and the ROM 43 is a CPU
It stores 44 control programs. The serial communication IC 40 and the serial communication driver 41 are for transmitting control data such as setting parameters from the CPU 44 to the various correction systems 47, and the I / O interface 4
Reference numeral 5 denotes a CPU for outputting an ON / OFF signal to various correction systems 47, inputting an ON / OFF signal from a sensor, and exchanging an ON / OFF signal with a system controller 48. The serial communication driver 46 exchanges data between the CPU 44 and the system controller 48.

The CPU 44 controls the CCD drive clock generation circuit 38, the sample timing control circuit 39, and the bus control systems 34 and 37 to take in the image data of the registration deviation measuring pattern output on the transfer belt and set the image position address. After confirming and calculating the registration shift amount, the serial communication IC4
0, through serial communication driver 41, or I /
Various correction systems are controlled through the O interface 45 and the serial communication 46.

Next, a concrete sample and correction algorithm will be described. 11 to 15 are views for explaining the algorithm of the rough adjustment pattern sample, and FIG.
To FIG. 19 are diagrams for explaining the algorithm of the fine adjustment pattern sample, FIGS. 20 to 21 are diagrams for explaining the algorithm of the sub-scanning sample start point correction, and FIG.
FIG. 6 is a diagram for explaining a correction algorithm of each color address error for K after sampling.

FIG. 11 shows the sampling of the coarse adjustment pattern.
As shown in (4), the pattern writing is started (step S51), the light amount is corrected and the shading is performed (steps S52 to S53), and the sampling start / end addresses of the K data in the main scanning direction are set (step S5).
4).

Then, the CPU waits until the K data sample end interrupt is generated (step S55), and transfers the sampling data (K data) in the main scanning direction to the main RAM in blocks (step S56).

After setting the sampling start / end addresses of the Y data in the main scanning direction (step S57), the image position of the K data in the main scanning direction is calculated (step S5).
8).

Wait until the Y data sampling end interrupt is generated (step S59), and the same process as in FIG.
As shown in FIGS. 2 to 13, Y data, M data, and C data are processed (steps S60 to S68). Then, as shown in FIG. 13, sampling data (C
After block transfer of (data) to the main RAM (step S69), a sample start / end address of Y data in the sub-scanning direction is set, and then an image position of C data in the main scanning direction is calculated (step S70).

Then, similarly to the sampling in the main scanning direction, the CPU waits until a K data sampling end interrupt is generated (step S71), and as shown in FIGS. Each process of block transfer of data, M data, and C data and address setting is performed (steps S72 to S84).

The above processing is repeated until it is completed three times (steps S85 to S88), and when it is completed three times, the image position of the C data in the sub-scanning direction is calculated (step S8).
9) The average value of the sampling data of three times is calculated (step S90).

In the fine adjustment pattern sampling, as shown in FIG. 16, after the pattern writing is started (step S101), the light amount is corrected and the shading is performed (steps S102 to S103) to sample the K data in the sub-scanning direction. A start / end address is set (step S104).

Then, the CPU waits until the K data sampling end interrupt is generated (step S105), and transfers the sampling data (K data) in the sub-scanning direction to the main RAM in blocks (step S106).

Subsequently, the sampling start / end address of Y data in the sub-scanning direction is set (step S107), and the sampling start / end address of K, Y, M, C data in the main scanning direction is further set.
After setting the end address (step S108), the image position of the K data in the sub-scanning direction is calculated (step S10).
9).

As shown in FIG. 17, the main scanning sample end interrupt is waited for (step S110), the sampling data in the main scanning direction is block transferred to the main RAM (step S111), and the main scanning image position is calculated. Is started (step S112).

Next, the CPU waits until a Y data sampling end interrupt occurs (step S113), transfers the sampling data (Y data) in the sub-scanning direction to the main RAM in a block manner (step S114), and then M in the sub-scanning direction.
Data sampling start / end addresses are set (step S115), and the image position of Y data in the sub-scanning direction is calculated (step S116).

Next, it is checked whether or not the calculation of the main scanning image position is completed (step S117), and if it is not completed, the calculation of the main scanning image position is continued (step S11).
8), as shown in FIG. 18, wait until an M data sample end interrupt is generated (step S119). In the same manner, the C data is processed as shown in FIGS. 18 to 19 (steps S120 to S131), and the process returns to step S105 until the sampling for the specified number of times is completed, and the same process is repeatedly performed. When the sampling is completed (step S132), the main scanning sample color (K → Y → M → C) is set (step S13).
3) Perform average calculation of sampling data (step S
134).

In the sub-scanning sample start point correction, as shown in FIG. 20, first, the nominal design sample address of each color is set (step S141), the sample completion is waited (step S142), and the image position of each color is calculated (step S142). Step S143). The same process is repeatedly performed for K, Y, M, and C until the sample is completed (step S144).

Next, the deviation amount Δ of the K image position address from the center of the previous K sample range is calculated (step S145). If the image position address cannot be confirmed due to dirt or the like in the previous sample, the correction value used before the second before is used.

The next K sample start and end addresses are calculated and set from (design value-deviation amount Δ) (step S
146-S147). Then, it waits for the completion of K sample (step S148).

Next, the image position of K is calculated as shown in FIG. 21 (step S149). And each color (Y, M,
The sample start / end addresses of C) are set (step S150), and the sample completion is waited (step S15).
1). K-Y, Y-M, and MC are constant values. This eliminates the need for correction when calculating the average value of the center-of-gravity addresses, and reduces the number of calculation steps. Next, each color (Y, M,
The image position of C) is calculated (step S152).

Step S until sampling of Y, M and C is completed
The process from step 15 is repeated (step S153), and the process from step S145 is repeated until the sampling is performed the specified number of times (step S154).

In the correction of the address error of each color with respect to K after sampling, as shown in FIG. 22, the pattern sample of each color (step S161), the image position address calculation (step S162), and the image position address average value calculation (step S163) are sequentially performed to correct Y, M, and C image position address average values (No. 1) (Image position address average value)-(KY, YM, MC) Correction value of error caused by fixing spread interval (design (Fixed value)) is performed (step S164).

Further, Y, M, and C image position address average value correction (No. 2) (image position address average value)-(ROS writing / CCD
The correction value (design fixed value) of the error caused by the read frequency mismatch is performed (step S165).

Then, the image position address error of Y, M and C with respect to K is calculated (step S166). That is,
Relative values are managed by YK, MK, and CK. Then, the correction value is calculated (step S167).

Next, the sampling system and the sampling method of the registration deviation measuring pattern of the present invention will be described. FIG. 23 is a diagram showing a configuration example of an image arithmetic circuit, and FIGS. 24 and 25 are diagrams for explaining a time domain for arithmetically processing image data.

The sampling system and the sampling method of the registration deviation measuring pattern are as follows: the CPU directly reads the image signal of the registration deviation measuring pattern from the CCD, or
There is also a method of switching to a mode in which the bus line is not used, releasing the CPU bus line to the CCD, or using a FIFO.

However, in the method of directly reading the image signal by the CPU, the high speed of the CPU is extremely pursued. Switching the mode without using the CPU bus line, CP to CCD
In the method of releasing the U bus line, it is necessary to stop the arithmetic processing of the CPU during sampling, and the operating rate of the CPU is extremely reduced. Further, in the method using the FIFO, there arises a problem that the address management of the image position signal of the sample pattern is very troublesome.

Therefore, according to the present invention, when the CCD samples the pattern output from the pattern generator, it comprises the main scanning high-speed image memory 35 and the sub-scanning high-speed image memory 36 shown in FIG. 10 by using the CCD clock. The data is written directly in the dedicated high-speed RAM. C
The PU 44 controls the bus control systems 34 and 37 so that the image data written in the main scanning high-speed image memory 35 and the sub-scanning high-speed image memory 36 is written into the work area of the main RAM 42 when they are not accessed. Move to and perform arithmetic processing.

When sampling in the sub-scanning direction, CC
When the image data captured from D is stored in the sub-scanning high-speed image memory 36, the sub-scanning image arithmetic circuit 33 calculates an average value of 16 lines, for example. Therefore,
As shown in FIG. 23, the sub-scanning image arithmetic circuit 33 is operated by AD.
An average value calculation circuit using DER may be used, but if the number of populations used for calculating the average value, that is, the number of pixels to be averaged is 2 to the nth power (n is a natural number), an average without error is obtained. You can easily find the value. The average value may be similarly calculated by the main scanning image arithmetic circuit when the image data fetched from the CCD is stored in the main scanning high-speed image memory 36 during sampling in the main scanning direction.

When the registration deviation measuring pattern as shown in FIG. 3 is used, the density of the registration deviation measuring pattern in the sub-scanning direction is high. As shown in 24, during the sample calculation processing SKP, SYP, SMP, ..., Sample calculation processing of the registration deviation measuring pattern in the main scanning direction is performed. Sample calculation processing SKP, SYP, SMP, SC of the registration deviation measurement pattern in the sub-scanning direction at that time
The sample calculation processing FKP1, FKP2, FKP3, and FKP4 of P and the registration deviation measurement pattern in the main scanning direction are
The previous calculation process must be completed by the time when the next sample calculation process SKP 'of the registration deviation measurement pattern in the sub-scanning direction and the sample FYS of the registration deviation measurement pattern in the main scanning direction start. Therefore, as shown in the figure, the calculation processing FKP1, FKP of the registration deviation measurement pattern in the main scanning direction is performed.
KP2, FKP3, and FKP4 are, so to speak, C after the calculation processing of the registration deviation measurement pattern in the sub-scanning direction is completed and before the calculation processing of the registration deviation measurement pattern in the sub-scanning direction is completed.
This is done by using the spare time for the PU.

The same applies to the use of the time domain when using a plurality of CCDs, and the CC on one side when the image output from the two CCDs is processed by one CPU.
FIG. 25 shows an example of how to use the time domain in D.
Is. In this case, the same method of using the time domain is performed for another CCD on one side, and the calculation processing time of the CCD on the other side is allocated to the vacant time for the calculation processing on the CCD on the other side to overlap them. Efficient use of the time domain can be implemented.

As described above, by allocating the high-speed image memory dedicated to the image data to the main scanning and the sub-scanning of each CCD, the absolute address of the high-speed image memory and the image address data can be easily compared. Furthermore, the image data written in the high-speed image memory is always processed by the CPU before the data is written in the same high-speed image memory before the data processing is completed. , The entire area of the high speed image memory can be allocated. Therefore, the management of the image address data can be made very simple.

When a plurality of CCDs are used to detect a registration deviation, a clock for moving each CCD is, for example, a registration deviation detection and correction board that handles CCD data on a board near the CCD. It is possible to reduce the number of circuits and the number of clocks operating in the registration shift detection and correction substrate by forming them on one substrate and distributing and supplying them to each CCD.
Therefore, the cost can be suppressed and the electromagnetic noise can be prevented from being inadvertently emitted.

When the CCD samples the pattern,
After calculating the average value using the ADDER circuit,
By directly writing to the high-speed image memory without passing through, the number of populations used for calculating the average value, that is, the number of pixels to be averaged is 2 to the nth power (n is a natural number) The CPU processing time can be greatly reduced.

Next, a sampling method for improving the correction accuracy will be described. The sample start timing necessary and sufficient for sampling the registration deviation measurement pattern is determined by the time when the registration deviation measurement pattern reaches directly below the CCD sensor from the position where the image forming apparatus outputs the registration deviation measurement pattern. The time is obtained from the speed of the transfer belt and the timing at which the image forming apparatus outputs the registration deviation measuring pattern, and the sampling interval is determined from the magnitude of the expected deviation.
Then, in order to always sample these registration displacement measurement patterns well, it is necessary to take a long sample interval. Therefore, the interval between the ROS output patterns of the registration shift measuring pattern also becomes long.

Therefore, in the present invention, first, a registration deviation measuring pattern having a wide interval as shown in FIG. 2 is used, and furthermore, each ROS output pattern is sampled widely at least once. Is used to find the address of each ROS output. Then, the DC component shift is corrected by correcting each ROS using the difference as the shift of each ROS. However, since various AC component deviations are ignored in sampling, an error corresponding to the maximum AC component deviation remains in some cases. In order to remove these various AC component deviation errors, a registration deviation measuring pattern having a narrow interval as shown in FIG. 3 is used, and moreover, it is several times sufficiently narrow with respect to the sample spread width of the coarse adjustment pattern. Desirably, one round of the transfer belt is repeatedly sampled. Then, similarly to the above, by adopting the center of gravity method or the like, the address of each color pattern is obtained and averaged once, and conversely, the write timing of each ROS is corrected so that the color registration deviation of each color is eliminated.

Note that, for example, the registration shift is determined by the first ROS (K).
When correcting the registration deviation of other ROS output based on the output,
The first ROS (K) may not have a correction function.
In this case, there is a problem that even if correction is performed after rough sampling, sampling cannot be performed when fine sampling is performed. Therefore, after the coarse sampling, the deviation of the first ROS (K) output is monitored, and the registration deviation measurement pattern is measured by the CCD.
By correcting the time to reach the position right below the sensor, it is possible to eliminate fine sampling error. In this way, not only coarse sampling → correction → fine sampling → correction, but also coarse sampling → correction → coarse sampling → correction → fine sampling → correction, coarse sampling → correction → fine sampling →
The correction accuracy, the fine sampling, and the correction may be combined in accordance with the correction accuracy and the target accuracy.

Further, the method of setting the sampling start point of the CCD will be described in detail. 26 to 27 are views for explaining the method of setting the sample start point of the present invention, FIG. 28 is a view for explaining another method of setting the sample start point of the present invention, and FIG. 29 is for measuring the registration deviation. It is a figure which shows the other example of a pattern.

The sample starting point in the CCD is
As described above, it is set based on the pitch between the image forming apparatus that forms the registration deviation measurement pattern and the CCD sensor, and the time from the output timing of the interface board to the point directly below the CCD sensor. The sampling start timing necessary and sufficient for sampling is set. In addition, when several registration displacement measurement patterns are sampled for each image forming apparatus in order to improve the image position address accuracy, the distance (pattern interval) from the previous sample pattern to the next sample pattern and the transfer belt conveyance The next sample start timing is a fixed time after the previous sample start timing based on the speed and the like.

However, due to the difference between the main scanning read frequency of the CCD and the main scanning writing frequency of the ROS and the existence of errors in the mechanism and its drive system, it is possible to always sample these image position recognition patterns well. The problem arises that you can't. Therefore, in the present invention, a method is adopted in which the sampling start timing of the registration deviation measurement pattern is made variable, and furthermore, these are sequentially corrected.

That is, as shown in FIG. 26, the first ROS
As for the sampling timing of the first pattern output from the first pattern, the time from the output timing of the interface substrate to the position directly below the CCD sensor is set by the pattern interval and the transfer belt conveyance speed. Then, the sampling timing of the output pattern of the second ROS is set from the distance from the output pattern of the first ROS to the output pattern of the second ROS, that is, the pattern interval and the transport speed of the transfer belt. Below 3rd ROS, 4th R
The same applies to the sampling timing of the output pattern of the OS.

Further, in order to improve the accuracy of the image position address, after the registration deviation measuring patterns of the first ROS to the fourth ROS are sampled, a plurality of blocks are sampled with these as one block and the average is taken for each image forming apparatus. ing. In this case, the sampling timing of the output pattern of the first ROS for the second time is obtained from the definite value of the image position address obtained by the centroid method or the like from the sampling result of the output pattern of the first first ROS as shown in FIG.

Further, as shown in FIG. 27, the sampling start point of the next pattern may be determined based on the definite image position address of the immediately preceding pattern, or as shown in FIG. The sampling start point of the next pattern may be determined based on the definite image position address of. A registration deviation measuring pattern as shown in FIG. 29 may be used.

By adopting the sampling method as described above, it is possible to always perform sampling in good timing without losing sight of the registration shift measuring pattern from the spread range, and not only the spread width itself can be narrowed but also absolute address management can be performed. It becomes unnecessary, and even a very small address counter can be used. Further, as a result, the amount of image data can be reduced, so that the capacity of the memory can be reduced and the calculation processing time of the CPU can be reduced. Therefore, the number of samplings per unit time can be increased, and higher AC components can be sampled.

Further, by carrying out for each pattern output from one image forming apparatus, the number of times of spread point correction can be reduced and the number of man-hours for correcting the fixed image position after sampling can be reduced, so that the processing amount of the CPU is reduced. It can be reduced. By having the correction value, the image position address can be corrected after sampling by a very simple calculation. Further, in order to remove the AC component, the sampling is performed a plurality of times for averaging, and when the averaging is used as the image position data, those average values are calculated and then the average image position address data is used only once for the image. It is only necessary to correct the position address. Therefore, by keeping the distance between the spread points constant and determining the image position with the correction value of the image position address that is held in advance, the C
It is possible to reduce the calculation man-hour of PU.

On the other hand, even when the image position cannot be determined from the sampling result due to scratches or stains on the transfer member, for example, dropout or dropout of the print medium such as toner, the deviation from the ideal sampling position is largely lost. Since a series of operations can be continued without doing so, even if the condition of the device deteriorates, it is possible to prevent immediate loss of function.

FIG. 30 is a diagram showing an example of a registration deviation measuring pattern having a heading pattern, FIGS. 31 and 32 are diagrams showing an example of setting a sample start reference timing of the registration deviation measuring pattern, and FIG. 33 is a heading. It is a figure which shows the other example of the pattern for registration shift measurement which has a pattern.

As a method of determining the sampling timing of the registration deviation measuring pattern, a heading pattern may be used as shown in FIG. In this case,
As shown in FIG. 1, the cueing pattern is sampled and the reference value of the registration deviation is measured from the time of returning from the image level to the background level by using the level specified value. Then, t1, t2, t3 shown in FIG. Sampling of the registration deviation measuring pattern is started after a prescribed time of t4. Further, as a measure against noise, a delay specified value (b) may be provided as shown in FIG. 32 to shift the sample start reference timing of the registration deviation measurement pattern. In the above example, the heading pattern is provided for each color, but as shown in FIG. 33, any one may be used for the outputs of a plurality of image forming apparatuses.

FIG. 34 (a) is a diagram showing another example of the registration deviation measuring pattern in the main scanning direction, and FIG. 34 (b) is a diagram showing another example of the registration deviation measuring pattern in the sub-scanning direction. .
The pattern shown in FIG. 34A may be used for measuring the registration deviation amount in the main scanning direction, and the pattern shown in FIG. 34B may be used for measuring the registration deviation amount in the sub-scanning direction.

By using such a pattern,
In the sampling result of the registration deviation amount in the sub-scanning direction, since a plurality of patterns are read at once, there is a difference between the design target value and the actual value, as in the case where the paper transport scanning moves slowly or faster than the target. Even in this case, sampling can be performed without being affected by the difference in moving speed, and the influence of magnification can be reduced. Furthermore, even if there is uneven vibration in the moving speed of the paper transport device,
The error is not absorbed in the sampling result of the registration deviation amount in the main and sub-scanning directions, and only the color registration component of the pattern itself can be sampled.
The influence of extra error factors can be excluded. Therefore, it is possible to eliminate an error factor at the time of sampling the registration shift amount due to the fluctuation of the moving speed of the easy transporting device, enable the optimum sampling, and obtain a reliable color registration shift amount.

35 to 37 are views showing an example of the seam detection system of the transfer belt, FIGS. 38 and 39 are views for explaining the operation of the seam avoidance timer, and FIG. 40 is a view for explaining the operation of the seam avoidance sensor. 41, FIG. 41 is a block configuration diagram of the registration deviation detection / correction system, FIG. 42 is a diagram for explaining image disturbance, FIG. 43 is a diagram showing an example of image average value standard value or more, and FIG. 44 is an image. FIG. 45 is a diagram showing an example below the minimum width standard value, and FIG. 45 is a diagram showing an example above the image width maximum standard value.

In FIGS. 35 to 37, 72K, 72
Y, 72M and 72C are writing devices of the image forming apparatus, 7
3K, 73Y, 73M and 73C are image forming devices such as photoconductor drums of the image forming apparatus, 74 is an image sensor, and 75
Is a light source, 76 is a transfer belt, 77A is a drive roller, 77
b is a driven roller, 78 is a sensor, 79 is a seam portion, 80
Reference numerals A and 80B denote reflecting devices (holes).

As shown in FIGS. 35 to 37, the transfer belt 7
6, a seam portion 79 is formed at the joint of the belt-shaped transfer member, and the distorted pattern image formed in this portion and the image forming unstable region in the vicinity thereof is erroneously recognized as a normal image. That is, when the registration deviation measurement pattern is sequentially sampled at a predetermined sample timing, even if there is a distorted pattern image formed in the image formation unstable area, if it is processed as it is, There is a problem in that there is a deviation with respect to the correction amount that should be. Therefore, in the present invention, the distorted pattern image formed in the image formation unstable region is prevented from being erroneously recognized as a normal image.

As the method, for example, as shown in FIG. 35, a sensor 78 detects a reflecting device (hole) 80A to detect that the seam portion 79 has come to a certain position in the upstream direction of the image pickup device 74. Image sensor 7
When 4 is sampling the pattern output from the image output device, the pattern and the seam portion 79 of the transfer belt 76 are overlapped with each other, or the data while they are close to each other are discarded, for example, for two blocks. Is. As a result, it is possible to reduce the influence of the AC component and improve the reliability of the center-of-gravity address data when the total average is taken.

Specifically, for example, as shown in FIGS. 38 and 41, the timers A and B are activated by the signal of the sensor 78, and the image formation is unstable between the end of the count of the timer A and the end of the count of the timer B. The data may be recognized as a region and the data may be discarded. Sensor 78, reflector 80
Depending on the position of A, only one timer may be used as shown in FIG. 39 and the timer may be activated by the signal of the sensor 78 and the time until the end of counting may be recognized as an image formation unstable region. The signal of the sensor 78 may be recognized as an image formation unstable region as shown in FIG.

In the processing of image data, the CPU discards the contents of the image memory containing the data sampled in the image forming unstable region, performs the calculation, and finally averages the calculation results. By not using the calculation result from the data sampled in, or by stopping the sampling until the image forming unstable region passes, it is possible to not sample in the image forming unstable region.

In detecting the unstable image forming area, the position of the reflecting device 80B is set to the image output devices 72K, 72Y, 72M, 72C, 73K, 73Y, 7 as shown in FIG.
If the detection can be performed corresponding to the timing at which the pattern is formed by the 3M and 73C, the image output signal can be controlled by the detection signal to stop the image output.

This prevents the distorted pattern image formed in the seam portion 79 of the transfer member and the image forming unstable region in the vicinity thereof from being mistakenly recognized as a normal image,
It is possible to prevent deterioration of the image position recognition accuracy.

Further, when adjusting the light quantity of the light source of the CCD or when performing shading correction, some points are taken in a state in which no pattern is formed, that is, an image is not formed and the copying member is idle. The CCD data is sampled with such a light amount, and the light amount for obtaining the best contrast among them is determined, or the shading correction amount is determined based on those values. At that time, if the image formation unstable region is sampled, erroneous light amount adjustment and shading correction will be performed. Therefore, in the present invention, the image forming unstable region is detected by the above method, the light amount of the light source of the CCD is not adjusted in the image forming unstable region, or the background data used for shading correction is not sampled. By doing so, erroneous light amount adjustment and shading correction value determination in the transfer member transmittance defective region due to dirt or scratches are prevented.

In each of the above-mentioned methods, the image forming unstable area is detected by using a sensor, a timer or the like. However, the registration deviation measuring pattern formed in the image forming unstable area is recognized from the sampled image profile. Then, by making these data invalid data, erroneous recognition as a normal image may be prevented.
Recognition of the registration deviation measuring pattern formed in the image formation unstable region is performed by, for example, image distortion as shown in FIG.
3 is performed by detecting data having an image average value standard value or more, data having an image width minimum standard value or less as shown in FIG. 44, and data having an image width maximum standard value or more as shown in FIG. You can

By doing so, it is possible to prevent the image of the distorted registration deviation measuring pattern formed in the seam portion and the image forming unstable region near the seam portion from being mistakenly recognized as a normal image. . As a result, it is possible to always give a correction amount with a small error to the registration deviation correction system, and it is possible to provide an image quality with very little registration deviation. Further, by using one sensor output, it is possible to recognize the seam portion of the transfer member and the range of the defective transmittance region of the transfer member due to stains or scratches in the vicinity of the seam portion. It is possible to stabilize the correction.

An image position address reading method of the present invention will be described. FIG. 46 is a diagram showing an example of a read image profile of a registration deviation measuring pattern, and FIGS. 47 to 51 are diagrams showing examples of an image profile with noise.

An ideal image profile when the registration shift measuring pattern is read is generally as shown in FIG. As described above, the center of this pattern image is found by using the center of gravity method, and the accurate image position address can be determined by repeating this operation and finding the average. If it becomes thick as shown by 47, the accurate image position address will not be obtained even if the center is obtained.

Therefore, in the present invention, an image pattern such as a stain is detected from the image level and the image width. For example, a threshold value is set as shown in (a) of FIG. 47, and an image width equal to or less than this threshold value is defined as a maximum image width standard value as shown in (b) of FIG. It is judged that the accurate image position address cannot be obtained due to stains or the like if it is provided.

Similarly, in the case of scratches, lack of toner, etc.,
Since the read pattern image becomes thin as shown in FIG. 48, it can be detected that the image width is equal to or less than the image width minimum standard value of (b). When the light amount of the light source itself is deteriorated or the transmittance of the transfer belt is deteriorated, the maximum standard value of the image level is exceeded as shown in FIG. 49. In the case of toner fog or the like, as shown in FIG. It can be detected as being less than the minimum standard value of the pattern image average value. Further, when the image width is equal to or less than the threshold value and a plurality of image width information is obtained as shown in FIG. 51, the image width is compared with the maximum image width standard value and the minimum image width standard value, When the data passes, it is not adopted.

When the read pattern image itself is far from the ideal image as described above, it is not used as data for determining an accurate image position address, so that an accurate image position address can be determined. . Further, since only the ideal image of the read pattern image is separated and used as data for determining the image position address, the registration error can be detected accurately by determining the accurate image position address even when there is a disturbance. It is possible to improve the quality of a color image by accurately correcting the registration shift.

[0108]

As is apparent from the above description, according to the present invention, the pattern generator uses a pattern for measuring coarse registration deviation or a pattern for measuring fine registration deviation in the main scanning direction and the sub-scanning direction of each image output means. Since it can be repeatedly generated and is selectively output from the image forming apparatus to perform sampling, the sample width can be set while adjusting the sample timing to increase the sample / adjustment efficiency and highly accurate registration deviation correction. It can be carried out. Moreover, the main scanning read frequency of CCD and RO
Even if the error is accumulated due to a difference in the main scanning writing frequency of S or an error in the mechanism or its drive system, the CCD loses sight of the pattern output from the image forming apparatus within the spread range. It becomes possible to perform efficient sampling, reduce the amount of image data to be captured, reduce the memory capacity, and shorten the calculation processing time. As a result, the number of samplings per unit time can be increased, and even AC vibration components can be sampled.

Further, since a dedicated memory is used when the CCD samples the registration deviation measuring pattern output from the image forming apparatus, high speed sampling is possible. Furthermore, since the processing of the CPU is distributed, the processing time of the CPU can be smoothed to reduce the amount of processing per unit time, the number of samplings per unit time can be increased, and even the AC vibration component can be sampled. You can

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of an embodiment of a registration registration image sampling correction method of a multiplex image forming apparatus according to the present invention.

FIG. 2 is a diagram showing an example of a pattern for an initialization cycle.

FIG. 3 is a diagram showing an example of a registration correction cycle pattern.

FIG. 4 is a diagram for explaining an algorithm for registration shift correction.

FIG. 5 is a diagram for explaining a registration adjustment algorithm during assembly and SKEW coarse correction.

FIG. 6 is a diagram for explaining an algorithm of a rough adjustment cycle.

FIG. 7 is a diagram for explaining an algorithm of a fine adjustment cycle.

FIG. 8 is a diagram for explaining an algorithm of a registration check cycle.

FIG. 9 is a diagram illustrating a configuration example of a registration registration image sampling correction device.

FIG. 10 is a block diagram showing a configuration example of a registration deviation correction substrate.

FIG. 11 is a diagram for explaining an algorithm of a rough adjustment pattern sample.

FIG. 12 is a diagram for explaining an algorithm of a rough adjustment pattern sample.

FIG. 13 is a diagram for explaining an algorithm of a rough adjustment pattern sample.

FIG. 14 is a diagram for explaining an algorithm of a rough adjustment pattern sample.

FIG. 15 is a diagram for explaining an algorithm of a rough adjustment pattern sample.

FIG. 16 is a diagram for explaining an algorithm of a fine adjustment pattern sample.

FIG. 17 is a diagram for explaining an algorithm of a fine adjustment pattern sample.

FIG. 18 is a diagram for explaining an algorithm of a fine adjustment pattern sample.

FIG. 19 is a diagram for explaining an algorithm of a fine adjustment pattern sample.

FIG. 20 is a diagram for explaining an algorithm for sub-scanning sample start point correction.

FIG. 21 is a diagram for explaining an algorithm of sub-scanning sample start point correction.

FIG. 22 is a diagram for explaining a correction algorithm of each color address error for K after sampling.

FIG. 23 is a diagram illustrating a configuration example of an image calculation circuit.

FIG. 24 is a diagram for explaining a time domain in which image data is arithmetically processed.

FIG. 25 is a diagram for explaining a time domain in which image data is arithmetically processed.

FIG. 26 is a diagram for explaining a method of setting a sample start point according to the present invention.

FIG. 27 is a diagram for explaining a method of setting a sample start point according to the present invention.

FIG. 28 is a diagram for explaining another method for setting the sample start point according to the present invention.

FIG. 29 is a diagram showing another example of the registration shift measuring pattern.

FIG. 30 is a diagram showing an example of a registration deviation measuring pattern having a heading pattern.

FIG. 31 is a diagram showing an example of setting a sample start reference timing of a registration deviation measurement pattern.

FIG. 32 is a diagram showing an example of setting a sample start reference timing of a registration deviation measurement pattern.

FIG. 33 is a diagram showing another example of a registration deviation measuring pattern having a heading pattern.

FIG. 34 is a diagram showing another example of the registration deviation measuring pattern in the main scanning direction and the sub-scanning direction.

FIG. 35 is a diagram showing an example of a seam detection system for a transfer belt.

FIG. 36 is a diagram showing an example of a seam detection system for a transfer belt.

FIG. 37 is a diagram showing an example of a transfer belt seam detection system.

FIG. 38 is a diagram for explaining the operation of the seam avoidance timer.

FIG. 39 is a diagram for explaining the operation of the seam avoidance timer.

FIG. 40 is a diagram for explaining the operation of the seam avoidance sensor.

FIG. 41 is a block diagram of a registration shift detection / correction system.

FIG. 42 is a diagram for explaining image distortion.

[Fig. 43] Fig. 43 is a diagram illustrating an example of an image average value standard value or more.

FIG. 44 is a diagram showing an example of an image width minimum standard value or less.

FIG. 45 is a diagram showing an example of an image width maximum standard value or more.

FIG. 46 is a diagram showing an example of a read image profile of a registration shift measuring pattern.

FIG. 47 is a diagram showing an example of a noisy image profile.

FIG. 48 is a diagram showing an example of an image profile with noise.

FIG. 49 is a diagram showing an example of an image profile with noise.

FIG. 50 is a diagram showing an example of an image profile with noise.

FIG. 51 is a diagram showing an example of a noisy image profile.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Control part, 2 ... Sampling control part, 3 ... Sampling part, 4 ... Data storage part, 5 ... Arithmetic processing part, 6 ... Pattern generator, 7 ... Image control part, 8 ... Image output part, 9
… Transfer belt

Front page continued (72) Inventor Hideaki Ashikaga 2274 Hongo, Ebina City, Kanagawa Prefecture Fuji Xerox Co., Ltd.

Claims (9)

[Claims] 1. A registration image sampling correction method for registering a multiple image forming apparatus, wherein a plurality of image output means are arranged along a transfer belt to form a multiple image, and each image output means is provided with a registration image sampling correction method. An output control unit that supplies image data and controls image formation, and a pattern generator that repeatedly generates a coarse registration deviation measurement pattern or a fine registration deviation measurement pattern in the main scanning direction and the sub-scanning direction of each image output unit, Sampling means for sampling the image formed by each image output means on both sides of the transfer belt in the main scanning direction, sampling control means for controlling the sampling start point and sample width of the sampling means, and the data sampled by the sampling means. Arithmetic processing that performs arithmetic processing to calculate registration error Means, data storage means for storing sampling data or arithmetic processing data, and control means for controlling image output and image sampling correction, and the control means performs rough adjustment when controlling image sampling correction. The registration deviation measurement pattern generated by the pattern generator is selected according to the mode or the fine adjustment mode, and the sampling start point and the sample width of the sampling control means are set to repeatedly generate the registration deviation measurement pattern to generate sampling data or A registration image sampling and correction method for a multiple image forming apparatus, characterized in that it is configured to integrate arithmetic processing data to obtain a pattern position. 2. The data storage means comprises a dedicated memory for storing image data of a registration deviation measuring pattern in the main scanning direction and a dedicated memory for storing image data of a registration deviation measuring pattern in the sub-scanning direction. The registration image sampling and correcting method of the multiple image forming apparatus according to claim 1. 3. The multiplex apparatus according to claim 1, wherein the arithmetic processing means preferentially processes the image data having a shorter sampling period and divides the remaining image data to disperse the processing. Registration sampling image correction method for image forming apparatus. 4. The multiplex processing apparatus according to claim 1, wherein the arithmetic processing means corrects the reference position of the registration deviation measuring pattern for each sample start point and integrates the image data with a deviation from the reference position. Registration sampling image correction method for image forming apparatus. 5. The arithmetic processing means identifies whether or not the image data is a registration deviation measurement pattern, and determines the processed data and the exclusion data based on the identification result of the registration deviation measurement pattern. 1. The registration-matching image sampling correction method of the multiple image forming apparatus described in 1. 6. The registration image sampling correction method for a multiple image forming apparatus according to claim 1, wherein the control means corrects the sampling start point based on the image position information of the previous pattern. 7. The control means corrects the sample start point based on the image position information of the previous pattern for each cycle while keeping the interval between the sample start points of the registration deviation measurement pattern constant. A registration registration image sampling correction method of the multiple image forming apparatus according to claim 6. 8. The control means selects a rough registration deviation measuring pattern in the coarse adjustment mode, and controls so as to capture sampling data during one revolution of the photosensitive drum of the image output means. 7. A registration registration image sampling correction method for a multiple image forming apparatus according to claim 6. 9. The control means selects a fine registration displacement measuring pattern in the fine adjustment mode and controls so as to capture sampling data during one rotation of the transfer belt. Registration image sampling and correction method for multiple image forming apparatus.