JP2790611B2 - Register error detection method and apparatus in multi-color printing press and register adjustment automatic controller - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and a device for detecting a register error in a multi-color printing press, and an automatic register adjustment control device.

In a multi-color printing press, a method of accurately detecting a register error of a register mark printed by each printing unit, a device for accurately detecting a register error, and a control of register adjustment for eliminating the detected register error. There is a need for a register adjustment automatic adjustment device that performs the following.

[0003]

2. Description of the Related Art In a multi-color printing press, a desired color image cannot be obtained unless the printing position of each color is matched. Therefore, an error in the printing register for each color is detected, and the registration is corrected to eliminate the error. Adjustments are being made.

A register error detecting method and apparatus for detecting a register error in multicolor printing and a register adjustment automatic control device include a type for detecting a printed register mark,
There is a type that detects a register mark on a printing plate mounted on a plate cylinder.

Japanese Patent Application Laid-Open Nos. Sho 63-22651 and Sho 62-2 disclose prior art which discloses a technique of detecting a printed register mark to which the present invention is directed.
No. 34934, JP-A-62-231755,
JP-A-58-217362 and JP-A-58-217362.
No. 20,457.

Japanese Unexamined Patent Publication No. Sho 63-22651 discloses a method in which a diamond mark printed by each printing unit is one-dimensionally scanned at a plurality of positions, whereby pixel data substantially corresponding to two-dimensional scanning is obtained. After the pixel data is sequentially loaded into the memory, the coordinates of the upper and lower vertices of each mark are calculated, the coordinates of the center of the mark are calculated from the coordinates of the upper and lower vertices, and the calculated coordinates of the center are already registered. This device calculates a registration error by comparing the coordinates with specified coordinates, and adjusts the registration of the top, bottom, left and right in order to eliminate the error. The coordinates of the vertices are determined by averaging.

Japanese Unexamined Patent Application Publication No. 62-234934 discloses a method of detecting a registration error in a top-to-bottom direction. The marks are arranged side by side in the direction of travel of the traveling paper, and the marks are sequentially and electrically detected by one detector in accordance with the traveling of the traveling paper. Is calculated, and the calculated distance is compared with a predetermined distance.
This is a method of calculating the registration error in the vertical direction.

Japanese Unexamined Patent Publication (Kokai) No. 62-231755 discloses a method for detecting a register error in the right and left direction. The marks are arranged side by side in the direction of travel of the running paper, and the marks are sequentially photoelectrically detected by one detector in accordance with the running of the running paper. In this method, a difference is obtained, and a registration error in the left-right direction is obtained from the difference.

Japanese Unexamined Patent Publication (Kokai) No. 58-217362 discloses a comparative example in which a plurality of lines of reference colors having a predetermined width and arranged in parallel at predetermined intervals are arranged in parallel so as to have a predetermined relationship. A predetermined print field in which a plurality of lines of color are overprinted is provided at a plurality of positions apart for each comparative color, and each print field is irradiated with electromagnetic energy (usually, visible, infrared, or ultraviolet light); The amount of energy reflected from each print field is sensed to determine the proportion of the non-print area of each print field, and from that proportion, the direction perpendicular to the line of each print field, ie, the top-to-bottom direction or the left-right direction. The error is calculated, and the difference in the ratio of the non-printing area between the print fields of the same comparative color at the separated positions is calculated. From the difference, the skew of the body, that is, the twist direction is estimated. Determining a difference, to eliminate register errors calculated, vertical direction, lateral direction, and a device to perform the register adjustment twist direction.

Japanese Unexamined Patent Application Publication No. 58-20457 discloses a misregistration inspection apparatus. Using an ITV camera, two bar-shaped marks are stored in a memory from an image pickup signal, and then the X coordinates of the two bars, that is, X1, X
From (2), (X2−X1) / 2 = K is calculated, and the set coordinate is B, and BK = ΔX is registered.

[0011]

Problems to be Solved by the Invention JP-A-63-2265
The technique disclosed in Japanese Patent Publication No. 1 (1995) has a problem that the accuracy of register error detection is reduced when the printing speed is increased or decreased, and thus the accuracy of register adjustment is reduced. That is, one-dimensional scanning by the CCD is started with a signal related to the phase of the plate cylinder rotation, and the scanning is continuously repeated every μs based on a clock signal, and the number of scans is counted.
The rotation circumference of the plate cylinder during the continuation of scanning is divided by the number of scans to obtain a pitch for each scan, and the value is used as the pitch in the running direction (top and bottom direction) of the running paper when the CCD detects the register mark. Since the coordinates are used to calculate the coordinates, for example, in newspapers, if the printing speed increases / decreases by 10,000 copies / hour per second, the running speed of the running paper at this time increases or decreases by 760 mm / s, which is the scanning cycle. A fluctuation of about 20 μm occurs in 26 μs, which causes a problem in that the coordinates lack accuracy and the accuracy of detection of registration error in the vertical direction and the accuracy of registration adjustment are reduced.

Each scanning is performed in one dimension,
Scanning is repeated a plurality of times for each register mark, and an amount of pixel data substantially equivalent to two-dimensional scanning is acquired.
Since this is processed by arithmetic processing to detect the registration error, a corresponding processing time is required and the load on the arithmetic means is large.
There is also a problem.

Furthermore, marks printed on newspapers or the like containing a lot of fibrous material cause image spots, that is, missing print dots, but there is no means for correcting this, and the detection accuracy is reduced. There is also.

Japanese Patent Application Laid-Open Nos. 62-234934 and 62-231755 require a register mark composed of two triangles, which inevitably increases the size of the register mark. However, there is a problem that the demand for a register mark as small as possible is contrary to a request for a register mark as small as possible due to a request for expansion of a printing surface.

Further, the spot diameter of the light beam irradiated on the paper must be reduced, and the detection accuracy cannot be increased. That is, there is a problem that the detection accuracy must be 10 μm or less, and an expensive device is required.

Further, since the light beam is applied, the photoelectric device for improving the detection efficiency must be arranged close to the running paper, which is disadvantageous for paper threading and maintenance work. is there.

Japanese Unexamined Patent Publication (Kokai) No. 58-217362 discloses an overlap printing (register mark) for detecting a registration error in the vertical direction and a parallel line for detecting a registration error in the horizontal direction. Overlap printing (register mark)
Therefore, it is necessary to separately provide print fields for these register marks, as disclosed in JP-A-62-234934 and JP-A-62-23.
There is a problem that it is contrary to the demand of the user, similarly to the one disclosed in Japanese Patent No. 1755.

In addition, since the amount of electromagnetic energy (usually visible, infrared, or ultraviolet light) applied to the overprinted field is reflected from the field, a registration error is obtained, so that the printing error is obtained. It is necessary to select the frequency of the electromagnetic energy to be used in consideration of the color to be printed and the ground color of the substrate, and the actual handling is complicated.
In some cases, it is necessary to provide a plurality of electromagnetic energy irradiation means and / or sensing means.

In Japanese Patent Application Laid-Open No. 58-20457, the method of calculating the center coordinates of the left and right of the mark is simply obtained as の of the bar size. It does not disclose whether it is an average or whether it is calculated from pixels in one horizontal scanning period. The same is true for the center coordinates of the top and bottom, and it is not clear. Even with this method,
As in Japanese Patent Application Laid-Open No. Sho 63-22651, marks printed on newspapers have spots on the image, that is, missing print dots. However, there is no means for correcting this, and the detection accuracy is reduced. is there.

Further, although the method of calculating the deviation is shown only for two bars, a large number of marks, for example, C,
No method is disclosed for calculating the deviation of the four bars M, Y, and Bk. Therefore, there is also a problem that it cannot be used for register adjustment for finding a deviation for four color bars.

The present invention solves the above-mentioned problems and can use a relatively small register mark, and prints not only on high-quality paper such as coated paper but also on rough paper such as newsprint despite the short detection time. A register error detecting method and apparatus for a multi-color printing press and a register adjustment automatic control device which have high register error detection accuracy even for printed matter, and therefore have high register adjustment accuracy and can reduce waste paper. The purpose is to do.

[0022]

In order to achieve the above object, the present invention is configured as follows. A register error detection method in the multi-color rotary press, the traveling paper, the run
The register marks of a rectangle or square and printed on each plate cylinder in a direction transverse to the row direction, the light source is a flash by reference pulse encoder output that rotates in synchronization with the rotation of the plate cylinder, synchronized thereto Read each register mark by matrix detection means and detect matrix
The matrix data of each register mark read by the means
Data of the preceding row line,
The binary image data of the succeeding row line and one column
The image data in the two row lines.
To repair defective reading using a logical filter
By repeating the above, register mark reading
Repair, and calculate the barycentric coordinates of each register mark from the read matrix data of each register mark, find the deviation between the barycentric coordinates of each register mark and a predetermined reference coordinate, and calculate this deviation. It is configured to be a register error.

[0023]

[0024]

The multi-color printing a misregister detection apparatus in a rotary press, the rotation and the encoder that outputs a reference pulse that rotates in synchronization of each plate cylinder, the traveling paper, each plate for each cylinder in a direction transverse to the traveling direction a light source for irradiating register marks printed rectangle or square, the luma trix detecting means to detect the register marks, and outputs a flashing start signal of the light source based on the reference pulses the encoder output, and flash detection start signal generating means for outputting a detection start signal of the matrix detector, Matricaria
Matrix of each register mark read by the
Of binary data of the preceding row line
Image data and subsequent row line binarized image data
Is shifted one column at a time, and the image in two row lines
Defective reading of data by logical filter
By repeating the repair of the
Image restoration means for repairing a mark read defect, barycentric coordinate calculating means for calculating barycentric coordinates of each register mark from matrix data of each register mark read by each matrix detecting means, and barycentric coordinates of each register mark And a deviation calculating means for calculating a deviation from a predetermined reference coordinate.

[0026]

[0027]

In a multi-color printing press equipped with plate cylinder phase adjusting means for adjusting the phase of the plate cylinder of each printing unit, an encoder for outputting a reference pulse rotating in synchronization with the rotation of each plate cylinder, and a light source for irradiating a registered mark of rectangle or square printed on each plate cylinder in a direction transverse to the running direction, and luma preparative <br/> helix detecting means to detect the register marks, the encoder output A flash detection start signal generating means for outputting a flash start signal of the light source based on the reference pulse to be output and a detection start signal of the matrix detection means;
The matrix data of each register mark
That is, following the binarized image data of the preceding row line
The row line binarized image data is serially
The image data in the two row lines
Repeatedly repairing defective reading parts using a filter.
By reading back, register mark reading failure
Image restoration means for performing restoration, barycentric coordinate calculating means for calculating barycentric coordinates of each register mark from matrix data of each register mark read by each matrix detecting means, and barycentric coordinates of each register mark; A deviation calculating unit for calculating a deviation from the reference coordinates; and an adjustment signal output unit for outputting an adjustment signal to a plate cylinder phase adjusting unit for adjusting a plate cylinder phase of the printing unit based on the deviation calculated by the deviation calculating unit. The configuration is as follows.

[0029]

[0030]

In the register error detecting method for a multi-color printing press according to the present invention, the register mark is formed in a rectangular or square shape, and pixel data of the register mark is read and detected by a rectangular or square detecting element of the matrix detecting means. . And the register
Modify the image by making the mark a rectangle or square.
Restoration is easy.

The image signal of the detected register mark is
A / D conversion is performed to obtain multi-valued data, which is binarized by a predetermined threshold value for each register mark. For the binarized image data, the barycenter coordinates of each register mark are calculated from the binarized image data over all the row lines.

Then, a deviation between the barycentric coordinate of each register mark and a predetermined reference coordinate is obtained, and this deviation is defined as a registration error. Before calculating the barycentric coordinates of each register mark, the binarized image data follows the binarized image data of the preceding row line in the matrix data of each register mark read by the matrix detection means. The row line binarized image data is
It is also possible to repair a register mark read failure by repeating a shift by one column and repeating the repair of the defective read portion by the logical filter for the image data in the two row lines.

In the register error detecting device for a multicolor printing press according to the present invention, the flash detection start signal generating means includes:
When the register mark reaches the position to be detected due to the travel of the traveling paper, a flash start signal of a light source is output based on a reference pulse output by an encoder that outputs a reference pulse that rotates in synchronization with the rotation of each plate cylinder. At the same time, a detection start signal of the matrix detection means is output, and the matrix detection means performs a reading operation.

The image signal of the detected register mark is
A / D conversion is performed to obtain multi-valued data, which is binarized by a predetermined threshold value for each register mark. The binarized image data is sequentially output to the barycentric coordinate calculation means for each row line.

The barycentric coordinate calculating means calculates barycentric coordinates of each register mark from the binarized image data over all the row lines, and outputs the result to the deviation calculating means. The deviation calculating means obtains a deviation between the coordinates of the center of gravity of each register mark and predetermined reference coordinates. In the present invention, this deviation is defined as a registration error.

When the binarized image data is sequentially output to the barycentric coordinate calculating means for each row line, the binary image data of the preceding row line is read out of the matrix data of each register mark read by the matrix detecting means. The shift of the digitized image data and the binarized image data of the succeeding row line by one column, and the restoration of defective reading portions by the logical filter for the image data in the two row lines is repeated, thereby making the register mark. Can be repaired.

Further, by providing the display means, the deviation calculated by the deviation calculation means can be displayed. The display means can display an alarm when the register mark cannot be detected in a predetermined state and when the calculated deviation is larger than a predetermined value.

In the automatic register adjustment control device for a multicolor printing press according to the present invention, the adjustment signal output means adjusts the phase of the plate cylinder based on the deviation of the register mark calculated by the deviation calculation means. By outputting the adjustment signal to the means, the register adjustment operation can be automatically performed.

[0039]

FIG. 1 is a schematic block diagram of an embodiment of the present invention. In FIG. 1, reference numeral 1 denotes a BB type printing unit, which forms a tower type printing press stacked in four stages. From below, Bk (ink), C (indigo), M (red), Y
Multicolor printing can be performed with the four colors (yellow). 2 is the plate cylinder,
3 is a blanket cylinder, 4 is an encoder that rotates in synchronization with the rotation of each plate cylinder, 5 is a web stock (running paper), 6,
7, 8, and 9 are guide rollers, and 10 is a folder. 11 and 12 are CCD cameras, 13 and 14 are motors and potentiometers for controlling the vertical direction of the divided cylinder, 15 and 16 are motors and potentiometers for controlling the horizontal direction of the divided cylinder, 17 is a register adjustment control panel, and 18 is an operation panel And a register adjustment display.

FIG. 2 is a detailed view of the CCD camera and the light source shown in FIG. 1, and shows a CCD camera constituted by a CCD matrix sensor and a xenon flash light source in relation to a register mark reading position. 11 and 12 are CCD cameras, 11 'and 12' are light sources, and 13 is a register mark printed on web stock.

FIG. 3 is a view schematically showing the mechanism of the divided plate cylinder. In the drawing, 19 is a motor with a speed reducer for controlling the plates on the C and D planes in the vertical direction, 20 is a potentiometer attached to the output shaft of the speed reducer, and 21 is the plate on the C and D planes in the horizontal direction. Is a potentiometer attached to the output shaft of the speed reducer. Reference numerals 23, 24, 25, and 26 denote a motor with a reduction gear and a potentiometer for controlling the printing plates on the side A and the side B in the top-to-bottom direction and the left-right direction. Reference numeral 27 denotes a printing plate mounted on the plate cylinder.

FIG. 4 is a diagram showing the positional relationship of register marks printed on web stock. 28 in the figure is 1
This shows a case in which a page is printed on a facing plate. Reference numeral 29 denotes a case where printing is performed on a facing plate having a two-page spread. Reference numeral 30 denotes a register mark printing area. Reference numeral 31 denotes a conventionally used register mark printing area. However, in the case of double-page spread printing, it becomes an image area, so that the conventional register mark cannot be printed there. Therefore, the conventional arrangement of register marks cannot be used, and the following arrangement of register marks is required.

FIG. 5 shows an example of the arrangement of register marks according to the present invention. Register mark is 1mm x 1mm
Within a rectangle or square. These register marks can be placed in the non-image area of the newspaper. Examples of register mark arrangement according to the present invention are (1),
(2), (3), (4), (5), and (6). Even if rectangles and squares are combined, the detection accuracy is
The same result is obtained.

Further, since the register mark is a rectangle or a square and the detection method is a method of detecting the barycentric coordinates, even if the image becomes thick or thin in printing, there is almost no influence on the detection accuracy. When the register mark is a triangle, a rhombus, or a circle, the edge of the mark is not accurately restored in the image restoration.

FIG. 6 shows a state in which the register mark is printed on newspaper. In the figure, 32 is a printed register mark, 33 is the actual round shape of the register mark, 34 is a state where the outer side is not printed,
Reference numeral 39 denotes a state in which print omission has occurred. As described above, when printing is performed on a newspaper surface containing a large amount of fibrous material, the printout of the register mark occurs, but occurs in an area of about 20 μm × 20 μm or less. The missing print refers to a state where the CCD sensor cannot detect the discrimination between the base color and the register mark when the amount of ink on the paper surface is small.

If a state occurs in which the acute angle portion of the mark is not printed, the accuracy of detecting the barycentric coordinates will be reduced unless these printing omissions are repaired. Although high-quality printing paper, such as coated paper, hardly causes print omission, the present invention is intended to be applicable to both newspaper printing and commercial printing.

Next, an example of the majority circuit for noise removal is shown in FIG.
Shown in Since the noise contaminates the paper surface due to the influence of ink mist generated during printing and causes fine mist to adhere to the paper surface, which hinders the reading accuracy of the register mark, the present invention employs a noise removal method. providing. The frequency of occurrence of these noises is usually about one or two in a continuous base color.

For the sake of explanation, FIG. 8 shows eight continuous image data. FIG. 8A shows that one noise is generated in the binarized image data output from the CCD camera. FIG. 8B shows that two noises are generated.

The image signal containing these noises is input to the noise elimination majority circuit shown in FIG. 7 as binary image data DT, and the pixel valid signal EN and the pixel shift signal CP are outputted.
As a result, the shift registers SR1 and SR2 operate, and the image signal is shifted. In this process, in the case of the image data shown in FIG. 8A, when noise is shifted to Q5 of the shift register SR1, the AND2 gate becomes HIGH, and this signal makes the OR1 gate HIGH, and the AND1 gate becomes HIGH.
Since 5 gates become HIGH, shift register SR
Clear 1 This operation removes noise. FIG.
In the case of the image data shown in (2), the shift register S
When Q5 of R1 becomes HIGH, AND3 gate becomes HI.
GH, and then OR1 (HIGH) → AND5
(HIGH) to clear the shift register SR1. This operation removes noise.

The timing chart of the above operation is shown in the timing chart of the majority noise elimination circuit (part 1) and the timing chart of the majority noise elimination circuit (part 2).
9 and 10.

FIG. 11 shows a simplified CC for explanation.
This shows a state in which the D matrix sensor has read one register mark. [1] indicates an image, and [0] indicates a non-image. H lines 0 to 17 indicate column lines, and V lines 0 to 17 indicate row lines. Image data [1] and non-image data [0] are serially output sequentially from row 0 in the column line direction.

FIG. 12 shows an example of a logical filter for image restoration. The logical filter is composed of four patterns of * 1, * 2, * 3 and * 4. In the case where the neighboring bits are formed as shown in the figure based on the illustrated [], the [X] bit is set to [1]. [] Is logic [1] as data. Data of preceding H line and subsequent H
An image can be restored by simultaneously shifting two lines of line data and passing through a logical filter.

FIG. 13 shows a state in which a logical filter is applied to image data. The row line (V
The data of lines 2 and 3 indicate the data of lines 2 and 3 in FIG. * 1 and * 2 at the missing position of the image
Applying the logical filter of, it can be understood that the data of the logical [0] is replaced with the logical [1], and the image is restored. (4,5), (6,7), (10,11), (1
2, 13) and (14, 15) indicate that the image can be restored in the same manner as described above. In order to perform image restoration, the register mark must have a rectangular or square shape as described above.

FIG. 14 shows an example of the image restoration circuit. FIG.
FIG. 5 shows an example of image data input to the image restoration circuit of FIG. This diagram is simplified for the sake of explanation, and shows image data composed of a preceding row line composed of 10 pixels and a succeeding row line composed of 10 pixels.

The operation of the image restoration circuit shown in FIG. 14 will be described below. 512x5 CCD matrix sensor
It is composed of 12 pixels, outputs serially the image data of the 0 row line, and then serially outputs the image data of the 1 row line, and sequentially outputs up to 511 row lines. When the image data of the preceding row line is input to the image restoration circuit, the data is stored in a 512-bit shift register SR. The image data of the preceding row line is serially output from the shift register SR in synchronization with the input of the image data of the succeeding row line.

FIG. 1 is a timing chart of the image restoration circuit.
6 is shown. This timing chart is created based on the case where the image data shown in FIG. 15 is input to the image restoration circuit.

In the image restoration circuit shown in FIG. 14, an image valid signal EN1, a binarized image signal DT1, and a clock pulse CP1 synchronized with the image are input to this circuit.
The image signal DT1 and the image valid signal EN1 are shifted in the respective registers in the order of R1 → R2 → R3, input to the shift register SR and stored.

Image valid signal EN1 and register R3
AND3 gate is OR1 until Q1 becomes LOW.
The shift operation of the shift register SR is performed by the shift clock CP1 pulse through the gate.

When the image signal of the next line is shifted to the register R1 in the same manner as described above, in synchronization with this shift,
The image signal of the preceding line stored in the shift register SR is shifted to the register R4 and the next CP
The signal is shifted to the register R5 by one clock, and further shifted to the register R6 by the next CP1 clock. This means that the preceding image signal passes through each register from R4 to R
5 → R6, and in synchronization with this, the subsequent image signal shifts each register from R1 → R2 → R3 in parallel.

In these states, the logical filter *
1 indicates that the logic of the register R5, the register R2 and the register R3 is [1], and the register R6
Performs the filter operation under the condition of logic [0]. Under this condition, the output of the AND7 gate becomes HIGH, and the OR
Since the output of gate 3 is HIGH, register R6
-Even if Q2 is LOW, the binarized image restoration signal DT2
The data is shifted by logic [1].

The logical filter * 2 includes the register R2
The logic of register R3 and register R5 is [1]
And the register R4 performs the filter operation under the condition of logic [0]. Under this condition, the output of the AND5 gate becomes HIGH, and the input in3 of the register R5 is input.
The register R5-Q3 becomes HIGH by the operation of the next clock pulse CP2. Subsequently, this signal is shifted to the register R6-Q3 by the shift clock pulse, and the register R6-Q is output by the OR3 gate.
Even if 2 is LOW, data is shifted to the binary image restoration signal DT2 by logic [1].

The logic filters * 3, * 4, * 5, and * 6 operate in the same manner as described above to restore image data by converting logic [0] to logic [1]. Can be.

The above processing is performed in real time 45 by hardware.
ms (printing speed 160,000 copies / hour) every time. Although a processor dedicated to image processing is commercially available, the image data from the CCD camera is temporarily stored in the memory for one frame, and then the filtering processing is performed to perform the arithmetic processing of the barycentric coordinates. The disadvantage is that it cannot be processed within When the detection method according to the present invention is used, high-speed processing can be performed, so that it is possible to reduce waste paper and improve detection accuracy, and to provide an inexpensive apparatus.

FIG. 17 shows an image detection area of the CCD sensor. The image detection area is 512 (H) x 5
It is composed of 12 (V) or more pixels. Figure 5 (1) shows the relationship between the image detection area and the reading position of the register mark.
2 shows the arrangement of the register marks shown in FIG.

The detection range of each register mark is defined by a horizontal line, the Bk mark is defined by XA, the C mark is defined by XB, the M mark is defined by XC, and the Y mark is defined by XD. The size of each mark is within 1mm x 1mm and 0.5mm x
It is 0.5 mm or more. Fig. 5 shows the arrangement of register marks.
In the detection range shown in (2), the horizontal line is defined by XA and XB, and the vertical line is defined by YA and YB.

Next, a method of detecting the center of gravity (center of gravity) of the register mark will be described. FIG. 18 shows a state where the CCD matrix sensor is reading one register mark. □ is logic [0] and the base color, ■ is logic [1]
Indicates a register mark, and corresponds to one pixel of the CCD sensor.

FIG. 18 shows 21 H lines and 21 V lines.
Although it is composed of lines, it is actually composed of 512 (H lines) × 512 (V lines). For the sake of simplicity, 21 (H line) × 21 (V line) was set.

The X barycenter coordinate is calculated as a value obtained by calculating the sum of the H line coordinate addresses of the image signal (register mark) and dividing by the number of pixels of the register mark. The Y barycenter coordinate is calculated as a value obtained by calculating the sum of the V line coordinate addresses of the image signal (register mark) and dividing by the number of pixels of the register mark.

FIGS. 19 and 20 show examples of the detection results of the X barycenter coordinate and the Y barycenter coordinate as an example of the register mark detection result (part 1) and an example of the register mark detection result (part 2).

The X barycenter coordinates and the Y barycenter coordinates are obtained from the sum of the pixels corresponding to the image read from the register mark, the sum of the X coordinate addresses corresponding to each pixel, and the sum of the Y coordinate addresses corresponding to each pixel. It can be obtained by the following equation.

X barycentric coordinates = (total of X coordinate addresses) / (total of pixels) Y barycentric coordinates = (total of Y coordinate addresses) / (total of pixels) The detected coordinates are the top and bottom directions of each mark with reference to the Bk mark. It is calculated by calculating the distance between the left and right directions. Then, a registration error is detected by detecting how much the detected coordinates deviate from the specified value. The reading position of the image is set so that the register mark on the paper surface to be printed is read when the register mark reaches a predetermined position.

FIG. 17 shows a state in which each register mark reading area is set in the image reading area of the CCD sensor in order to read the register mark as described above.

FIG. 21 shows an example of a barycentric coordinate calculation circuit.
Hereinafter, a method of calculating the barycentric coordinates will be described with reference to FIG.
The pulse generated from the encoder in synchronism with the rotation of the plate cylinder of the rotary press is an instruction pulse signal indicating the reference rotation angle position of the plate cylinder, which generates one pulse per rotation of the plate cylinder, and is generated in synchronization with the rotation of the plate cylinder. And a rotation synchronizing pulse signal. These pulses are input to I0 and I1 of the timing generation circuit. The timing generation circuit receiving these pulse signals generates a TO signal and an ST signal indicating that the printed register mark has reached the center of the CCD matrix camera reading area.

The signal input to I1 of the timing generation circuit indicates that the timing generation circuit
It is input to generate a signal indicating that the mark reading position of the D matrix camera has been reached.

The ST signal is input to the CCD matrix camera and gives a reading start signal. Upon receiving this signal, the CCD matrix camera resets the pixels and simultaneously starts exposure.

On the other hand, the TO signal is input to the flash control circuit, and causes the flash lamp F to flash at the same time when the CCD matrix camera starts exposure. C
The exposure time of the CD matrix camera has a shutter of about 1 μs, and the flash of the flash lamp F is also performed within 1 μs. These configurations are used in a state where a still image is obtained. With the above operation, the CCD matrix camera reads the register mark and sequentially sends out one frame of image signal.

The column line data valid signal EN, the field valid signal FE, the binarized image signal DT, and the image reading clock pulse signal CP are input from the CCD matrix camera to the noise removal majority circuit. The noise removal majority circuit removes noise from the image data,
A column line data valid signal EN1, a field valid signal FE1, a binary image signal DT1, and an image reading clock pulse signal CP1 are output. These signals are input to the image restoration circuit, and the image data is restored by the same processing as described for the operation of the image restoration circuit shown in FIG. The restored image data signal is transmitted to the center of gravity detection circuit.

Outputs Q1 to Q of CNT (X1) counter
Reference numeral 9 indicates the X coordinate address (address of the row line [horizontal line]) of the register mark. This CNT
(X1) The counter counts the image valid signal EN2 and the CP2 clock signal output in synchronization with the pixel. 51
It returns to 0 with the second BO signal.

The output of the CNT (X1) counter is input to the decoder DEC, and the output of the decoder DEC is
The quadrant divided area of the X coordinate shown in FIG. 17 is decoded and output. That is, a signal for dividing the horizontal line into four areas is output.

The reading area of the X coordinate of the Bk mark is CN
The output of the T (X1) counter is 0 to 127, the reading area of the X coordinate of the C mark is CNT (X1) counter output is 128 to 255, and the reading area of the X coordinate of the M mark is CNT (X1) counter output of 256 to The reading area of the X coordinate of the 383 and the Y mark has the CNT (X1) counter output defined by 384 to 511, respectively, and these areas are decoded and output to XA, XB, XC, and XD. For the Y coordinate, 0 in the above XA
Vertical lines 0 to 511 indicate the Bk mark area, vertical lines 0 to 511 in the XB indicate the C mark area, and XC
The vertical lines 0 to 511 indicate the M mark area.
Vertical lines 0 to 511 in D indicate the Y mark area.

RXA register, ADH1 addition controller,
And an accumulator composed of an HA register constitutes a circuit for accumulating the address of the X coordinate of the pixel corresponding to the mark in the XA area. An accumulator composed of the RXB register, the ADH2 addition controller, and the HB register constitutes a circuit for accumulating the address of the X coordinate of the pixel corresponding to the mark in the XB area. RXC register,
The accumulator including the ADH3 addition controller and the HC register constitutes a circuit for accumulating the address of the X coordinate of the pixel corresponding to the mark in the XC area. The accumulator including the RXD register, the ADH4 addition controller, and the HD register constitutes a circuit for accumulating the X-coordinate address of the pixel corresponding to the mark in the XD area.

RYA register, ADV1 addition controller,
And an accumulator composed of the VA register constitutes a circuit for accumulating the address of the Y coordinate of the pixel corresponding to the mark in the XA area. An accumulator composed of the RYB register, the ADV2 addition controller, and the VB register constitutes a circuit for accumulating the address of the Y coordinate of the pixel corresponding to the mark in the XB area. RYC register,
An accumulator composed of an ADV3 addition controller and a VC register constitutes a circuit for accumulating the address of the Y coordinate of the pixel corresponding to the mark in the XC area. An accumulator composed of the RYD register, the ADV4 addition controller, and the VD register constitutes a circuit for accumulating the address of the Y coordinate of the pixel corresponding to the mark in the XD area.

The accumulation of the X coordinate address of the register mark in the XA area is performed as follows. When the XA output of the decoder DEC is HIGH and the DT2 output (image) is HIGH, the AND1 gate operates by the CP2 image reading clock pulse, the clock pulse a is input to the RXA register, and the output of the CNT (X1) counter (X The coordinate address is latched by the RXA register. Next, the data stored in the HA register and the X address data latched by the RXA register are added by the ADH1 addition controller by the a 'timing pulse, and stored in the HA register again.

X in each area of XB, XC, and XD
The accumulation of the coordinate addresses is also performed by the circuit operation of the AND2 gate, the AND3 gate, and the AND4 gate in the same manner as the accumulation of the X coordinate addresses of the register marks in the XA area.

The accumulation of the Y coordinate address of the register mark in the YA area is performed as follows. When the XA output of the decoder DEC is HIGH and the DT2 output (image) is HIGH, the AND1 gate operates by the CP2 image reading clock pulse, the clock pulse a is input to the RYA register, and the output of the CNT (Y1) counter (Y (Coordinate address) is latched by the RYA register. Next, the data stored in the VA register and the Y address data latched by the RYA register are added by the ADV1 addition controller by the a ′ timing pulse, and the added data is stored in the VA register again.

The Y in each of the YB, YC, and YD areas
The accumulation of the coordinate addresses is performed in the same manner as the accumulation of the Y coordinate addresses of the register marks in the XA area by the circuit operations of the AND2 gate, the AND3 gate, and the AND4 gate.

The CNT (SA) counter counts the total number of pixels corresponding to the marks in the XA area. CNT (S
B) The counter counts the total number of pixels corresponding to the marks in the XB area. CNT (SC) counter is XC
The number of all pixels corresponding to the marks in the area is counted. CN
The T (SD) counter counts the total number of pixels corresponding to the marks in the XD area.

The CNT (SA) counter counts the total number of pixels corresponding to the marks in the XA area by counting the number of times the AND1 gate is operated. CNT (SB)
The counter counts the total number of pixels corresponding to the marks in the XB area by counting the number of times the AND2 gate operates. The CNT (SC) counter counts the total number of pixels corresponding to the marks in the XC area by counting the number of times the AND3 gate operates. The CNT (SD) counter counts the total number of pixels corresponding to the marks in the XD area by counting the number of times the AND4 gate is operated.

Next, the calculation of the barycentric coordinates will be described.
An image signal for one screen is transmitted from the CCD matrix camera, and data is stored in each register.

As described above, the accumulated value of the X coordinate address of the register mark in the XA area is stored in the HA register, and the Y coordinate address is stored in the VA register. The total number of pixels of the register mark in the XA area is held by the CNT (SA) counter.

The accumulated value of the X coordinate address of the register mark in the XB area is stored in the HB register, and the address of the Y coordinate is stored in the VB register. The total number of pixels of the register mark in the XB area is held by a CNT (SB) counter.

The accumulated value of the X coordinate address of the register mark in the XC area is stored in the HC register, and the Y coordinate address is stored in the VC register. The total number of pixels of the register mark in the XC area is held by the CNT (SC) counter.

The accumulated value of the X coordinate address of the register mark in the XD area is stored in the HD register, and the Y coordinate address is stored in the VD register. The total number of pixels of the register mark in the XD area is held by a CNT (SD) counter.

The image restoration circuit supplies the EN2 signal and CP2
The signal is input to I2 and I3 of the timing generation circuit. The timing generation circuit calculates 1 based on these signals.
It is possible to recognize that the image signal for the screen has been completely transmitted from the CCD matrix camera.

When the timing generation circuit recognizes the completion of image transmission, it generates a T1 pulse. With this T1 pulse, the MPX1 multiplexer and the MPX2 multiplexer operate, and the contents of the HA register and CNT
(SA) The contents of the counter are input to the divider GX. Next, division is performed by the Ta pulse, and the X barycentric coordinates in the XA area are calculated. At the same time, the MPX3 multiplexer operates by the T1 pulse, and the contents of the VA register and the contents of the CNT (SA) counter are input to the divider GY. Next, division is performed by the Ta pulse, and the Y barycenter coordinates in the XA area are calculated.

In the same manner as above, the T3 pulse and the Ta pulse are generated, and the coordinates of the X barycenter and the coordinates of the Y barycenter in the XB area are calculated. The T5 pulse and the Ta pulse are generated, and the X barycenter coordinates and the Y barycenter coordinates in the XC area are calculated. The T7 pulse and the Ta pulse are generated, and the X barycenter coordinates and the Y barycenter coordinates in the XD area are calculated.

The T2 pulse, the T4 pulse, the T6 pulse, and the T8 pulse are used for transmitting a deviation to a register error adjusting device described later. FIG. 22 shows a series of these timing charts as a timing chart of the center-of-gravity coordinate calculation circuit.

In the divider equation, GX = K · X / S (1) and GY = K · Y / S (2) where K is a constant and represents the pitch between the upper, lower, left, and right pixels. If the pitch between the pixels of the line and the pitch between the pixels of the vertical line are not equal, K
Take different values.

Next, the adjustment of the register error will be described.
FIG. 23 shows an example of the register adjustment control circuit. The output GX of the divider GX in FIG.
The XA register latches the X barycentric coordinates of the register mark in the XA area from O by the T2 timing pulse. At the same time, in order to adjust the registration error in the vertical direction, the YA register latches the Y gravity center coordinate of the register mark in the XA area from the output GYO of the divider GY in FIG. 21 by the T2 timing pulse.

Similarly, by the T4 timing pulse, X
The XB register and the YB register respectively latch the X barycenter coordinates and the Y barycenter coordinates of the register mark in the B area.

Similarly, the T6 timing pulse causes X
The XC register and the YC register respectively latch the X barycenter coordinates and the Y barycenter coordinates of the register mark in the C area.

Similarly, the T8 timing pulse causes X
The XD register and the YD register latch the X barycenter coordinate and the Y barycenter coordinate of the register mark in the D area, respectively.

Next, XA, XB,
On the basis of the X barycentric coordinates stored in the XC and XD registers and the reference coordinate data KXA, KXB, KXC, and KXD, each comparator COMP calculates the deviation, and the D / A converter Output as analog voltage. At the same time, YA, YB, YC and Y
Based on the X barycenter coordinate stored in each register of D and the reference coordinate data KYA, KYB, KYC, and KYD, each comparator COMP calculates the deviation, and calculates D / A
It is output as an analog voltage by the converter.

These analog voltages are compared with the voltage of the potentiometer PM by a differential amplifier, and are controlled by rotating a motor by a driver circuit so that the difference becomes zero. In this way, the top and bottom left and right registration errors are adjusted. If the deviation is zero, the motor does not rotate and no adjustment control is performed.

In addition to the above-described method of performing the adjustment control based on the deviation from each of the reference coordinates, the distance between the center of gravity of the Bk register mark and the center of gravity of the other register marks can be calculated based on the Bk register mark. It is also possible to adopt a method that is set in advance and adjusts the error of the distance. The reference mark is not limited to the Bk register mark, but may be any of the C mark, M mark, and Y mark.

In the figure, PM is a multi-rotation potentiometer to which a constant voltage is applied. The adjustment motor is
It is a motor with a reduction gear, and its output shaft and the potentiometer PM are linked.

As described above, the example in which the register adjustment control circuit is constituted by hard logic has been described. However, the register adjustment control circuit may be replaced by a microprocessor instead of the hard logic. CCD
The matrix camera is not limited to the above-described transmission timing, and may use a camera having a different output method. Also, a color CCD matrix camera can be used.

[0108]

The method and apparatus for detecting a register error in a multicolor printing press according to the present invention and the automatic register adjustment control apparatus have the following effects.

When the register mark reaches the reading position, a read start signal is given to the CCD matrix sensor by an external trigger to control the shutter so that the exposure time is constant, and the flash light source is caused to flash in synchronization with the shutter control. Therefore, the still image of the register mark can be read. Therefore,
The register mark can be read accurately without being affected by the printing speed. Thereby, the detection accuracy of the register error is improved.

Since the noise removal circuit is provided, the register error can be adjusted without being affected by the print stain at the time of printing, so that the waste paper can be reduced and the resource saving can be greatly improved.

[0111] By adopting a rectangle or square as the shape of the register mark, it is possible to easily repair an image of a partial print omission of the mark printed on the fibrous newsprint, so that the register mark can be accurately read. Can be. Thereby, the detection accuracy of the register error is improved.

Since noise removal and image restoration can be performed in real time when image data is transmitted from the CCD matrix camera, it is possible to detect a registration error on a paper surface printed at high speed. This can reduce waste paper.

If a telecentric lens is employed in the optical system of the CCD matrix camera, the influence of the front and back of the paper and the change in the magnification of the register mark due to the curvature of the guide roller can be neglected. Correction of register marks becomes unnecessary. This makes it possible to provide a simple logic configuration, so that an inexpensive and easily adjustable device can be provided.

Since the register mark is a rectangle or a square and the method of detecting the coordinates of the center of gravity is used, there is almost no effect on the detection accuracy of the thick or thin image in printing.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of an embodiment of the present invention.

FIG. 2 is a partial detailed view of FIG. 1, that is, a detailed view of a CCD camera and a light source, showing a positional relationship between a CCD camera constituted by a CCD matrix sensor and a register mark read by a xenon flash light source.

FIG. 3 is a view showing a schematic mechanism of a divided plate cylinder.

FIG. 4 is a diagram showing a positional relationship of register marks printed on web stock paper.

FIG. 5 is a diagram showing an example of register mark arrangement according to the present invention.

FIG. 6 is a diagram showing a state in which a register mark is printed on newspaper.

FIG. 7 is a diagram illustrating an example of a majority circuit for noise removal.

FIG. 8 is a diagram illustrating an example of image data.

FIG. 9 is a diagram showing a timing chart (No. 1) of a majority noise elimination circuit.

FIG. 10 is a diagram showing a timing chart (2) of the majority noise removal circuit.

FIG. 11 shows a simplified CCD matrix sensor 1
FIG. 10 is a diagram showing a state in which register marks are read.

FIG. 12 is a diagram illustrating an example of a logical filter for image restoration.

FIG. 13 is a diagram showing a state in which a logical filter is applied to image data.

FIG. 14 is a diagram illustrating an example of an image restoration circuit.

FIG. 15 is a diagram illustrating an example of image data input to the image restoration circuit.

FIG. 16 is a diagram showing a timing chart of the image restoration circuit.

FIG. 17 is a diagram showing an image detection area of the CCD sensor.

FIG. 18 is a diagram showing a state where a CCD matrix sensor is reading one register mark.

FIG. 19 shows an example of a detection result of a register mark (part 1).
FIG.

FIG. 20 shows an example of the result of register mark detection (part 2).
FIG.

FIG. 21 is a diagram illustrating an example of a barycenter coordinate calculation circuit.

FIG. 22 is a diagram showing a timing chart of the barycentric coordinate calculation circuit.

FIG. 23 is a diagram illustrating an example of a register adjustment control circuit.

[Explanation of symbols]

Reference Signs List 1 BB type printing unit 2 plate cylinder 3 blanket cylinder 4 encoder that rotates in synchronization with rotation of each plate cylinder 5 web paper (running paper) 6 guide roller 7 guide roller 8 guide roller 9 guide roller 10 folder 11 CCD camera 12 CCD camera 13 Motor and potentiometer for controlling the vertical direction of the divided cylinder 14 Motor and potentiometer for controlling the vertical direction of the divided cylinder 15 Motor and potentiometer for controlling the lateral direction of the divided cylinder 16 Motor for controlling the lateral direction of the divided cylinder and Potentiometer 17 Register adjustment control panel 18 Operation panel and register adjustment display

Claims (5)

(57) [Claims] 1. A method for crossing a traveling direction on a traveling paper
The register marks of a rectangle or square to print for each plate cylinder on the direction, the light source is a flash by reference pulse encoder output that rotates in synchronization with the rotation of the plate cylinder, by a matrix detector in synchronization with this Read each register mark, each register mark read by the matrix detection means
Of the preceding row line in the matrix data of
Binarized image data of the following row line and the binarized image data
Are shifted one column at a time, and
Image data reading failure due to logical filter
By repeating the repair of the part, the register
Repair the mark reading error, calculate the barycentric coordinates of each register mark from the read matrix data of each register mark, and calculate the deviation between the barycentric coordinates of each register mark and the predetermined reference coordinates. A register error detecting method in a multicolor printing press, wherein the deviation is set as a register error. 2. The method according to claim 1, wherein a majority logic filter is used as the logic filter. 3. A rotation and encoder that outputs a reference pulse that rotates in synchronization, the traveling paper, register marks rectangle or square printed on each plate cylinder in a direction transverse to the running direction of the plate cylinder a light source for irradiating a luma trix detecting means to detect the register marks, and outputs a flashing start signal of the light source based on the reference pulses the encoder output, flash light for outputting a detection start signal of the matrix detector Detection start signal generation means and each register mark read by the matrix detection means
Of the preceding row line in the matrix data of
Binarized image data of the following row line and the binarized image data
Are shifted one column at a time, and
Image data reading failure due to logical filter
By repeating the repair of the part, the register
Image restoration means for repairing a reading error of a mark, barycentric coordinate calculating means for calculating barycentric coordinates of each register mark from matrix data of each register mark read by each matrix detecting means, and barycentric coordinates of each register mark. And a deviation calculating means for calculating a deviation from a predetermined reference coordinate. 4. Oite to claim 3, whereas the deviation calculating means to display the calculated deviation, when the registered mark could not be detected in a predetermined state, and when the calculated deviation is excessive than a predetermined value And a display means for displaying an alarm. 5. A multi-color printing press including a plate cylinder phase adjusting means for adjusting a phase of a plate cylinder in each printing section, wherein an encoder for outputting a reference pulse rotating in synchronization with the rotation of each plate cylinder; to a light source for irradiating a register mark of rectangle or square printed on each plate cylinder in a direction transverse to the running direction, and luma trix detecting means to detect the register mark, the reference pulses the encoder outputs A flash detection start signal generating means for outputting a flash start signal of the light source based on the signal and outputting a detection start signal of the matrix detection means, and each register mark read by the matrix detection means.
Of the preceding row line in the matrix data of
Binarized image data of the following row line and the binarized image data
Are shifted one column at a time, and
Image data reading failure due to logical filter
By repeating the repair of the part, the register
Image restoration means for repairing a reading error of a mark, barycentric coordinate calculating means for calculating barycentric coordinates of each register mark from matrix data of each register mark read by each matrix detecting means, and barycentric coordinates of each register mark. A deviation calculating means for calculating a deviation from a predetermined reference coordinate; and an adjustment signal for outputting an adjustment signal to a plate cylinder phase adjusting means for adjusting a plate cylinder phase of a printing unit based on the deviation calculated by the deviation calculating means. An automatic register adjustment control device for a multicolor rotary press including an output unit.