JPH0147823B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to a method for accurately inspecting the printing quality of continuously running printed matter in a rotary printing press or the like while the machine is still running. Generally, very strict conditions are required regarding the print finish of printed matter, so a sufficiently reliable method is also required for inspection. Furthermore, in rotary printing presses and the like, the printed printing section runs at high speed, so it is impossible to inspect it in real time using a visual method. Therefore, as an inspection method in such cases, the image of the moving printed matter is captured using a one-dimensional image sensor camera, etc., the resulting image signal is grasped as an analog quantity, and it is compared with a reference voltage. A method has been proposed in which the quality of the pattern is determined by using the image signal, or the image signal is binarized using a predetermined threshold value to determine the quality of the pattern. However, although this method using analog image signals provides sufficient accuracy for detecting defects on plain originals, it is insufficient for determining pass/fail when the density changes over a wide range, such as in patterns on printed materials. The drawback was that accuracy could not be obtained. Therefore, after digitizing the image signal and making it possible to store the image signal over the entire pattern part of the printed matter to be inspected, the image signal obtained from the printed matter of standard finish is stored as reference data, and then the image signal to be inspected is A method has been proposed in which an image signal read from a printed matter is used as inspection data, and stored reference data is read out corresponding to each pixel of the inspection data and compared, thereby making a quality determination. Therefore, Figure 1 shows the principle of an example of an inspection method using this digitization method, where P is a printed matter, CY is a printing cylinder, 1 is an image sensor camera, and 2 is an analog-to-digital (A/D) converter. SW is a changeover switch, M is a reference data memory, and CD is a comparator. The printed matter P is a long piece of paper, film, etc., on which a predetermined pattern is successively and repeatedly printed along the running direction by a rotary printing machine, and sent out by a printing cylinder CY. An image sensor camera (hereinafter referred to as an IS camera) 1 images the surface of the printed material 1 on which the pattern is printed, and one-dimensionally captures a predetermined portion along the width direction X perpendicular to the running direction Y of the printed material P. Scan and extract image signals. Then, the one-dimensional image signal read from the printed matter P is digitized for each predetermined pixel by the A/D converter 2 and supplied to the switch SW. The reference data memory M has the storage contents as shown in FIG. 2, and is designed so that digitized density information for each pixel can be written and read at addresses a divided in the printing width direction and the running direction. It is configured. Therefore, if the switch SW is switched to the reference data memory M as shown in the figure, and a predetermined pattern part of the printed matter P is imaged by the IS camera 1, density information for each pixel in the part along the width direction are sequentially written to the addresses divided in the printing width direction of the reference data memory M, and this is sequentially repeated by the movement of the printed matter P in the running direction Y, and the density is written to the addresses divided in the running direction of the reference data memory M. The image data is written as information, and image data of a predetermined range of the picture printed on the printed material P is eventually written into the reference data memory M. When the switch SW is switched to the lower side opposite to that shown in the figure, the comparison circuit CO sets the image data sequentially read from a predetermined range of the pattern of the printed matter P as inspection data I, and correspondingly reads the image data sequentially for each pixel. The reference data S read out from the reference data memory M is received as the comparison input, respectively, and the comparison result is generated as the output J. Therefore, at a predetermined time immediately after starting the printing operation, after confirming that a printed matter with no defects has been obtained, the switch SW is switched to the reference data memory M, and at that time, the data obtained from the printed matter P is transferred. The image data read from the printed matter P is written in the reference data memory M as the reference data S, and if the switch 5 is then switched to the comparison circuit CO, the image data sequentially read from the printed matter P is stored one after another as the inspection data I in the comparison circuit CO. Each pixel is sequentially compared with the input reference data S read out from the reference data memory M, and the result is obtained as an output J. Therefore, from now on, by checking the output J of the comparator circuit CO and checking whether the reference data S and the inspection data I match, it is possible to continuously check the quality of printing of the printed matter P that is running at high speed during the printing operation. It is possible to perform the inspection with a reliability that is incomparably higher than that of visual inspection. The purpose of the present invention is to improve the conventional printed matter inspection apparatus using such a digitization method, to easily provide a highly accurate and effective masking function, and to be able to easily respond to changes in the state of printed matter. To provide a printed matter inspection device which can always provide stable operation even when the object is in use. In order to achieve this objective, the present invention not only compares reference data and inspection data pixel by pixel, but also compares the total sum of the entire pattern to be inspected and compares the data in a straight line along the running direction of the printed material. Compare the total sum of pixels arranged in
It is characterized by the fact that it makes a pass/fail judgment based on the overall half of these comparison results.Furthermore, it detects the running speed of the printed matter, the horizontal deviation of the pattern, etc. and rewrites the reference data, and also determines the comparison level for each pixel. Another feature is that it can be set arbitrarily. Embodiments of the printed matter inspection apparatus according to the present invention will be described below with reference to the drawings. FIG. 3 is a block diagram of the entire system in one embodiment of the present invention, showing the printing cylinders CY, IS1, A/
The D converter 2, etc. are the same as the conventional example shown in FIG.
RE is a rotary encoder provided in the printing cylinder D, 3 is a picture position detector, 4 is a picture position signal input circuit, 5 is a running position signal input circuit, 6 is a scanning direction signal input circuit, 7 is a digital input interface, 8 is an address generation circuit, 9 is a first feature extraction comparison judgment circuit, 10 is a second feature extraction comparison circuit, 11 is a third feature extraction comparison judgment circuit, 1
2 is a general judgment circuit, 13 is a reference data memory rewrite signal generation circuit, 14 is a reference data memory (corresponding to M in FIG. 1), 15 is a pixel control data memory,
16 is a buffer memory, 17 is a computer interface, 18 is a monitor address generation circuit, 19 is a monitor interface, 20 is a monitor, and 21 is a computer.
Connected via BUS. Now, as already explained, in such an inspection system, it is necessary to correlate the reference data with the inspection data from the pattern that is repeatedly printed for each pixel. You must decide on an address that will be uniquely represented. Therefore, a rotary encoder RE is used to detect the rotational direction position of the printing cylinder CY necessary for this addressing, and the signal is input to the traveling position signal input circuit 5, and an inspection on the circumference of the printing cylinder CY is started. Output signals representing the point and end point. FIG. 4 shows an embodiment of this running position signal input circuit 5. Here, the output signals from the rotary encoder RE are a ZERO signal that generates only one pulse at each predetermined rotational position each time the printing cylinder CY rotates once, and an A-phase signal that generates a predetermined number of pulses per rotation. There are two types. First, before starting the test, computer 2
1 to latch 22 via computer interface 17, which enables generation of the MEND signal to indicate the end of the test. In all of the circuits described below, setting values are written to the latches using the computer 21 in the same manner as described above. After being reset with the ZERO signal, the counter 24 counts the A-phase signal, and its output is compared with the set value of the latch 22 in the comparator 23. When the two match, the output is a monostable multivibrator (MMV). ) 26 and MEND
A signal is generated and an inspection end point is set. The inspection starting point is the rotary encoder RE.
This is represented by a ZERO signal, and this is called an MZERO signal. That is, in this embodiment, the signal MZERO and
The inspection period is between MEND. On the other hand, since the A-phase signal of the rotary encoder RE is used as a clock signal in other circuits, it is also supplied as is to the BUS, and is used as MCLK. Next, to detect the position in the print width direction required for addressing, we use a self-propelled IS1 (for example,
CCD camera, MOS camera, etc.) and use the output signal from it. Now, the scanning of IS1 is controlled by an external synchronizing signal shown in FIG. 5, and line scanning is repeated at predetermined scanning intervals only when this signal becomes High. As shown in Figure 5, IS1 generates a START signal at the start of scanning, and synchronizes with the scanning by using SCLK.
Generate a signal. The scanning direction signal input circuit 6 is a circuit that controls the input and output of the three types of signals described above, namely, the external synchronization signal, the START signal, and the SCLK signal, and one embodiment thereof is shown in FIG. First, the traveling position signal is supplied from the driving position signal input circuit 5.
An external synchronization signal is generated by the MZERO signal, which starts scanning IS1. Next, in order to repeat scanning at the same interval in the rotational direction of the printing cylinder CY, the MCLK signal is counted by a counter 27.
The comparator 28 compares the number of divisions in the rotational direction of the printing cylinder CY, which is preset in the latch 30.
When the count value reaches the set value, the comparator 28 generates an external synchronization signal. Repeat this for one inspection screen. On the other hand, when IS1 scanning starts, this circuit
A START signal and a SCLK signal are received from IS1, and a counter 32 counts the SCLK signals. When the number of divisions in the scanning direction set in advance in the latch 29 is reached, a comparator 31 is used to generate a SEND signal. Further, the SZERO and SEND signals are controlled using a D type flip-flop 37 and the MZERO and MEND signals so that they are generated only during the inspection period. Here, the inspection surface of printing cylinder CY and MZERO,
Figure 7 shows the relationship between the MEND, SZERO, and SEND signals. The address generation circuit 8 uses the outputs of the traveling position signal input circuit 5 and the scanning direction signal input circuit 6 to generate an address on one inspection screen. That is, as shown in Fig. 8, the printing cylinder
By dividing CY into n parts in the rotational direction and m parts in the scanning direction, the inspection screen is divided into _n 〓 ⁱ⁼¹ _o 〓 ^j=1 Aij addresses, and these addresses are generated. is shown in Figure 9. First, the counter 50 is cleared by the MZERO signal and the address is initialized. next,
By inputting the SZERO signal, an address in the scanning direction for the first line is generated using SCLK, and the number of addresses is counted by a counter 48. This part is composed of a D type flip-flop 43, an AND 44, and a counter 48. Further, the number of divisions in the scanning direction is set in latch 46 in advance. By comparing this set value with the count value by the comparator 49, the number of addresses for one scanning line is determined. Then, the address generated by the counter 50 is outputted from the BUS to the necessary circuits via the latch 51. Similar operations are repeated up to n lines. In this embodiment, an internal clock CLK is used to capture digital data in synchronization with the address, and a write signal WR is used in synchronization with the address to write the captured data into the reference data memory 14 and buffer memory 16. is occurring. In this embodiment, an A/D converter 4 and a digital interface 7 are used to input image information obtained by photoelectric conversion in the IS1 into the present apparatus. By the way, in a typical A/D converter, the output digital signal is linear with respect to the input analog signal. However, the density value D used by humans when identifying colors is expressed as, for example, the transmission density D=log ₁₀ I ₀ /I (I ₀ : amount of light at the time of incidence, I: amount of light after transmission), It has logarithmic characteristics with respect to the amount of light. Therefore, even in actual inspection equipment, it is preferable to take the logarithm of the photoelectrically converted signal because it is relatively close to the human sensory scale. Therefore, in this embodiment, as shown in FIG. 10, an A/D converter 4 is used which can obtain a digital signal from a photoelectrically converted analog input with both linear and nonlinear characteristics.
The structure is such that logarithmic characteristics can also be approximated.
Incidentally, the timing of the digital data output of this A/D converter 4 is outputted from the address generation circuit 8 in order to synchronize with the above-mentioned address. It uses CLK. Next, in this embodiment, it is possible to select between a method in which the digital data input to the digital interface 7 is taken into the present apparatus as it is, and a method in which the image is enhanced and then taken in as data. The latter method aims to emphasize the existence of defects by emphasizing each pattern on the printed surface, and can be realized by using the Laplacian of spatial filtering. It is shown in Fig. and Fig. 12. At this time, the latch 53 is set in advance to determine whether to use the digital data as it is or to use the image-enhanced data.
The signal selector 54 selects one of them and imports it into the device as data. The image enhancement circuit shown in Figure 12 uses spatial filtering.

It is composed of [formula]. That is, to find the density Dij of a certain pixel, use its four neighbors,
It is expressed as follows. Using this,
Dij is Dij=4D _ij −(D _i-1,j +D _i,j-1 +D _i,j+1 +D _i+1,j
)
becomes. At this time, if there are no constraints on the configuration,
It goes without saying that 8-neighborhoods, etc., can be used instead of 4-neighborhoods, which would lead to even better results. Now, in FIG. 12, shift registers 55, 56,
Using latches 57 to 65, adders 66, 67, 68, 71, 73 are added so as to satisfy the above equation.
A shift register 69, inverters 70 and 74, and exclusive OR 72 are used. In this embodiment, in order to synchronize with the address generated by the address generation circuit 8, shift control of each data is performed.
This is done using CLK. The configuration and signals of the data input system can be summarized as shown in FIG. 13. By the way, in this embodiment, computer 2
1 controls the operation of the entire system, but the internal state of this inspection device at this time is as follows:
It has a function. (i) Setting mode (ii) Reference mode (iii) Inspection mode (iv) Stop mode And these modes are control commands.
It can be changed using the Ccmmand signal. First, in (i) setting mode, setting values can be set for each circuit from the computer 21. For this reason, each circuit has a computer 2 as shown in FIG.
It has common components so that it can receive the same information. Here, the dip switch 79 is set with a value unique to each circuit. Now, to set the settings for these circuits,
The address CAddress2 signal that selects the circuit to be set from among the many circuits and the dip
It is necessary to check the match with switch 79. moreover
A signal from a decoder 77 that decodes the Command signal and makes the circuit recognize that it is in the configuration mode, a specific latch (e.g. 75) for data configuration in the circuit.
A total of 3 CAddress1 signals that can specify
The final data setting location can be determined using two signals. Data CDATA is written into the selected latch 75 by a signal CWR from the computer 21. In this way, in this embodiment, each set value can be easily set and changed using the computer 21. Next, when the printing operation is started and good printed matter is obtained, the switch is made to (ii) the reference mode, whereby the data inputted from the digital interface 7 is transferred to one screen and the address The data is written into the reference data memory 14 using the Address output from the generation circuit 8 and the write signal WR. From the next screen, the device automatically enters the (iii) inspection mode and compares and judges the data in real time with the data captured in the reference mode. The circuits that perform this comparison/judgment process are first, second, and third feature extraction comparison/judgment circuits 9, 10, and 11.
It is. Note that (iv) stop mode is used to stop the function of this inspection device. First, the first feature extraction comparison and determination circuit 9 compares and determines the reference data SD and the inspection data ID for each pixel at the same address. That is, as shown in FIG. 15, the data SDij read from the reference data memory 14 and the digital interface 7
Data can be entered in real time through IDij
Pixels with 1SDij−IDij1>determination level are extracted as defective pixels. FIG. 16 shows an example of this, in which data SD read out from the reference data memory 14 and data inputted from the digital interface 7 are shown.
ID is supplied to latches 82 and 83, and then calculated by adders 84 and 86, exclusive OR 85, and inverter 91, and the absolute value of the difference between SD and ID is written into latch 88. Thereafter, the comparator 90 compares the judgment level data CD1 read from the pixel control data memory 15 and set in the latch 89 in synchronization with the above data, and the result is sent to the D type flip-flop 87.
Each pixel is output as the determination result J1. Next, in the second feature extraction comparison judgment circuit 10,
As shown in FIG. 17, the absolute value of the difference between the total sum for one screen of reference data and the total sum for one screen to be inspected is determined and compared with the determination level. i.e. printing cylinder CY
When the number of divisions in the rotational direction is n, the number of divisions in the scanning direction of IS1 is m, the reference data is SDij, and the inspection data is IDij, then for these, _n 〓 ⁱ⁼¹ _o 〓 ^j=1 SDij and _n 〓 ^{i =1} _o 〓 ^j=1 IDij. This improves detection accuracy for low-density defects that spread over the entire screen and gradual density changes. FIG. 18 shows one embodiment of this, in which data SD read from the reference data memory 14 is transferred to a latch 94,
95, an adder 98, and an AND 92 for one screen, and written to the latch 102 using the MEND signal. Digital interface 7
Similar operations are performed on the inspection data ID taken in through the latches 96 and 97, an adder 99, and an AND93. However, since the ID is the difference between the two, it is written into the latch 103 through the inverter 105 in order to convert it into a one's complement number. In addition, in order to clear the addition data, use AND92 and 93 once per screen to close latches 94 and 93.
Enter 0 in 96. For this purpose, a 0 signal is generated once per screen using the MEND signal, the MZERO signal, and a D type flip-flop 100. Standard data SD calculated as above
The sum of all pixels of , and that of the inspection data ID are added to adders 106 and 108, and exclusive OR 10.
7. The signal is inputted to one input A of the comparator 110 through a circuit for calculating the absolute value of the difference, which is composed of an inverter 109. Here, the judgment level data CD read out from the pixel control data memory 15
2 and those exceeding the judgment level are D type flip flop 112 once per screen.
is output as a determination signal J2. Furthermore, the third feature extraction comparison and determination circuit
As shown in FIG. 9, the absolute value of the difference between the sum of the reference data SD in the traveling direction y of the printed matter P, that is, the rotational direction of the printing cylinder CY, and that of the inspection data ID is determined and compared with the determination level. That is, the reference data
This is to compare _o 〓 ^j=1 SDij and _o 〓 ^j=1 IDij for SDij and inspection data ID _ij . This improves the accuracy of detecting defects in the rotational direction (for example, so-called doctor streaks) that frequently occur in gravure prints and the like. FIG. 20 shows an example of this, in which the data SD read from the reference data memory 14 is AND12
0, an adder 121, a memory 127, and a transceiver 129 at the dividing point in the scanning direction.
The total sum _o 〓 ^j=1 SD _ij of each rotation direction is calculated. The memory 127 is updated with data added one scanning line at a time, and finally becomes data added in the rotational direction for one screen. On the other hand, similar operations are performed on the test data ID taken in through the digital interface 7 using the AND 122, the adder 123, the memory 131, and the transceiver 133. Memory 127, 128, 13
For access to 1,132, a D type flip-flop 113 and a counter 114 are used to generate an address for each scanning line.
In addition, as a method of calculating the results added in the rotational direction, data of the previous inspection screen is transmitted to the transceivers 130 and 134 while adding one line of the next inspection screen.
The data is written to the memories 128 and 132 through the process. A method is adopted in which the previous inspection screen is judged from the second line onwards. This enables continuous real-time processing. To obtain the above timing, D type flip-flop 118, OR119 and SEND,
MEND and CLK signals are used. moreover,
Writing the test data ID into the memory 132 is performed after converting it into a one's complement number using the inverter 140 because it is necessary to find the difference from the reference data SD later. Adder 1 adds the absolute value of the difference between the reference data SD obtained as described above at the scanning division point _o 〓 ^j=1 SD _ij and that of the inspection data ID _o 〓 ^j=1 ID _ij
35, 137, an exclusive OR 136, and an inverter 139.
Furthermore, using the comparator 138, it is compared with the judgment level data CD3 read out from the pixel control data memory 15, and if the data exceeds the judgment level, it is divided by the number m of division points in the scanning direction on one screen, and the D type flip. Judgment signal J3 from flop 125
is output as Next, the first, second, and third feature extraction/comparison/judgment circuits 9, 10, and 11 individually compare and determine and output the results, which are comprehensively judged as one detection screen, and are used to detect markers, alarms, or surrounding areas. A comprehensive judgment circuit 12 is a circuit that outputs a signal that causes an action to be taken to an output device. First, in this circuit, regarding the defect signal of the first feature extraction comparison judgment circuit 9 that compares each pixel, it is determined that the detected screen is defective only when the defect signal occurs a number of times equal to or higher than a certain judgment level. . That is, the determination level of defective products inspected by the inspection apparatus according to this embodiment can be changed according to the inspection performance required. On the other hand, the second and third feature extraction comparison and determination circuits 1
Defect outputs of 0 and 11 are considered serious defects, and if even one occurs on the detection screen, the entire screen is considered defective. FIG. 21 shows an example of this, in which the number of pixels, that is, the number of CLKs, is counted only when a defective signal is generated by inputting the judgment output J1 of the first feature comparison and judgment circuit 9 to the gate of the counter 142. That's what I do. This number of defects in one screen and latch 145
Comparator 146 compares the judgment level set in
Set using MEND signal. Next, the second and third feature extraction comparison judgment circuit 1
The judgment outputs J2 and J3 of 0 and 11 are supplied as they are to the D-type flip-flops 143 and 144, respectively, and when a defect signal is generated on the detection screen, the D-type flip-flop 14
3,144 is set immediately. Depending on the presence or absence of the above three signals, the overall judgment signal TJ on the detection screen is OR14.
It is output through 8. Therefore, by monitoring the overall judgment signal TJ from the overall judgment circuit 12, it is possible to inspect printed materials that are moving at high speed during printing operations with extremely high accuracy and in real time. Next, the picture position detector 3, picture position signal input circuit 4, and reference data memory rewrite signal generation circuit 13 will be explained. As described above, the inspection device according to this embodiment is mainly applied to rotary printing presses. By the way, the position of the pattern that is continuously and sequentially formed on printed matter by such a rotary printing machine in the direction perpendicular to the running direction of the printed matter, that is, in the printing width direction, is not always at the same position during the printing operation. . That is, the feeding operation of printed matter to a multicolor rotary printing press generally involves considerable widthwise fluctuations. Therefore, in such cases, the operator adjusts the printing register by moving the printing cylinder of the printing press in the width direction each time, or the adjustment is performed automatically.
Therefore, even if there is no abnormality in the printing operation and the pattern is printed correctly, the position in the width direction of the printed material is not necessarily the same, but always fluctuates. On the other hand, since the inspection apparatus according to this embodiment has a method of determining the quality of printing by comparing the inspection data and the data read from the reference data memory 14, it is possible to If the position of the pattern on the printed material P changes in the width direction from when the data is read, even if the pattern is correct, both data will no longer match, resulting in a malfunction. This will be explained in more detail with reference to FIGS. 22 and 23. In Fig. 22, ISC is an imaging part by an image sensor included in IS1 (Fig. 3), and has a configuration in which a large number of photoelectric conversion elements, for example, 512, are arranged linearly in the width direction of printed matter P. On the other hand, a to d represent patterns repeatedly printed in sequence on the surface of the printed matter P. These patterns a to d have different positions in the x direction, that is, in the width direction, of the printed matter P for the reasons described above. Next, in FIG. 23, one of the photoelectric conversion elements of the photoelectric conversion section ISC is indicated by ISC', and one of the pixels of patterns a and b are indicated by a' and b', respectively. Therefore, data is now written in the reference data memory 14 in the state of picture a in FIG. 22, and then,
Assume that data according to pattern b in FIG. 22 is read. As a result, the relationship between a certain pixel a' of pattern a and the photoelectric conversion element ISC' becomes as shown in FIG. 23, and the relationship between a certain pixel b' of pattern b and the photoelectric conversion element ISC' becomes as shown in FIG. Suppose that it becomes as shown in FIG. 23 2. Then, the density information written to the corresponding address of the reference data memory 14 will be the data corresponding to the area a'' in FIG. 23 3, while the inspection data will be the data corresponding to the area b'' in FIG. 23 4. I end up. Therefore, at this time, even if patterns a and b are the same, the data read from the reference data memory 14 and the inspection data will no longer match, and it will be determined that the printing is defective even though the pattern is correct. It will end up being put away. Therefore, a reference data memory rewrite signal generation circuit 13 is provided, and this circuit takes in the picture position signal from the picture position detector 3 and outputs the picture position signal to the reference data memory 14.
The pattern position data when the data is written is held, and thereafter, each time the inspection data is read, it is compared with the captured pattern position data, and when the difference between the two exceeds a predetermined value, a reference data memory rewrite signal is sent. is generated, and the image data read from the pattern at that time is written as new reference data. and,
For this purpose, the picture position detector 3 includes an image sensor, detects the position in the width direction of the picture printed on the surface of the printed matter P that is traveling by the printing cylinder CY, and sends it to the BUS via the picture position signal input circuit 14.
Sends a picture position signal to. Next, an example of a method for detecting a picture position using the picture position detector 3 is shown in FIG. In FIG. 24, a indicates a pattern printed on the printed material P, and 3' indicates a portion detected by the pattern position detector 3. The picture position detector 3 images a predetermined portion in the width direction x of the picture printing surface of the printed matter P that is continuously fed in the y direction by the printing cylinder CY, and detects the image of a predetermined portion in the width direction A detection signal is generated when a predetermined portion enters the detection portion 3'. Therefore, the pattern a on the printed matter P
When the position of changes in the width direction, that is, in the x direction, the pattern a
The size of the portion included in the detection portion 3' changes, and the amount of light incident on the picture position detector 3 changes. Therefore, by detecting the amount of light incident on the picture position detector 3, the position of the picture a in the x direction can be detected, so the light quantity detection signal from this detector 3 is taken as a picture position signal,
The signal may be sent to the BUS via the picture position signal input circuit 4. Next, FIG. 25 shows another embodiment of the picture position detection method, in which M is a wedge-shaped register mark. In multi-color rotary printing presses, wedge-shaped registration marks are placed in the margins of the printing surface for automatic registration in the width direction of printed matter. Therefore, in this embodiment, this register mark M is imaged by the pattern position detector 3, and the detected portion of the register mark M is A change in the area contained within ' is detected as a change in light amount, and a picture position signal is thereby obtained. Therefore, according to this embodiment, a mark suitable for position detection can be used regardless of the original pattern, so that the detection operation can be performed more reliably. Furthermore, FIG. 26 shows another embodiment of the method for detecting the position of a picture, in which N is a mark for detecting positional deviation printed in the margin of the printed matter. This mark N consists of a large number of line segments printed at regular intervals along the width direction x of the printed material in parallel to the running direction y, and the line included in the detection portion 3' depending on the position change Δx in the width direction. The pattern position signal is obtained by utilizing the change in the number of minutes. Therefore, according to this embodiment, it becomes easy to obtain the picture position signal as the number of pulses. Next, FIG. 27 shows an embodiment of the reference data memory rewrite signal generation circuit 13, in which input/output ports P ₁ to
Microcomputer (referred to as microcomputer) MC with P ₆ , latches 150 to 152, counter 1
53, and AD154. FIG. 28 is a flowchart explaining the operation. When the microcomputer MC starts operating and enters the operation according to this flow, the microcomputer MC first checks the content of the control command Command, and
When it is in "stop" or "set" mode, it repeats reading the control command Command. However, in the "setting" mode, position tolerance data CD is sent from latch 150 to port _P2 of microcontroller MC.
Enter 4. Also, when the control command Command is in the "reference" mode, the latch 152
The picture position information PD is taken into the microcomputer MC via _P6 , and this is stored as the reference position information of the picture at the time of taking in the reference value data, and then the process returns to reading the control command Command. Then, when the control command Command enters the "inspection" mode, the operation is as follows. The position tolerance data CD4 input to port _P2 in the "setting" mode is transferred to the register in the microcomputer MC. In the same manner as in the "reference" mode, the picture position information PD is taken in, and it is compared with the reference position information as the position information when the detection value data is taken in, and the absolute value of the difference is calculated. Next, the absolute value of this difference and position tolerance data
CD4 and when the absolute value of the difference exceeds the position tolerance, a reference data memory rewrite signal is sent.
Generate RWR. Return to reading the control command Command. Therefore, according to this embodiment, whenever the positional deviation of the pattern of the printed matter P in the width direction becomes large and exceeds a preset value, the reference value memory rewrite signal generation circuit 13 outputs a signal. The rewrite signal RWR is supplied to the reference data memory 14, and the image data being read by IS1 at that time is written to the reference data memory 14 as new reference data.
After that, the newly written reference data is used to determine the inspection data until the rewrite signal RWR is generated again. Even if a positional shift occurs in the width direction perpendicular to the running direction, no malfunction occurs, and only the quality of the pattern can be accurately determined. Next, the latch 151 and counter 153 in this reference data memory rewrite signal generation circuit 13,
The functions performed by AND154 and microcomputer MC will also be explained. As mentioned above, in the inspection apparatus according to this embodiment, as shown in FIG.
The image data is read by scanning in the direction x perpendicular to the running direction y of the printed matter P. Therefore, the image data actually read by IS1 is in the direction x of the running direction y of the printed matter P.
It is read along direction A, which is a composite of the scanning direction x by IS1. Therefore, when the traveling speed of the printed matter P changes,
The direction A in which image data is read is also, for example, a in FIG. 30.
It changes as shown in b. In other words, even if the scanning speed by IS1 is constant, the printed matter P
When the running speed of the vehicle increases, the direction A read by IS1 becomes B. As a result, when we look at the same pixel detection areas C and D of IS1 in Figures 30a and 30b, we see that in Figure 30a, data with a black pattern is read in the entire area, whereas in Figure 30b, data with a black pattern is read within the detection area. , some of the data of a white pattern is included, and even though the same pattern is read, the pixel data of the corresponding address becomes different. Therefore, if the running speed of the printed matter P differs between when the reference data is set and when the inspection data is read, an erroneous inspection result will be given if left as is. Therefore, the latch 151 is loaded with speed tolerance data CD.
5 is accepted, and BUS is sent from the rotary encoder RE to AND154 via the traveling position signal input circuit 5.
The RWR signal is generated when the running speed of the printed material P changes by more than a permissible value between when setting the reference data and when reading the inspection data. This operation will be explained using the flowchart shown in FIG. Now, if a signal is given to the circuit 13 via the BUS, the microcomputer MC first issues a control command.
Read the contents of the Command.
In the case of "stop" or "setting" mode, reading of the control command Command is repeated. however,
In the "setting" mode, speed tolerance data is sent to port _P5 in the microcontroller MC via latch 151.
CD5 is written. In addition, when the control command Command is in the "reference" mode, the microcomputer MC outputs a signal to reset the counter 153, resets the counter 153, gives a count start signal to the AND154, and outputs a pulse MCLK from the rotary encoder RE. take in. Microcomputer MC measures a predetermined time. The microcomputer MC outputs a count end signal to the counter 153 and the rotary encoder
Finish capturing pulse MCLK from RE. The count value of the counter 153 is input to the microcomputer MC via port _P3 , and this is input to the microcomputer MC.
It is stored in the register inside as the traveling speed information at the time of importing the reference value data. Return to reading the control command Command. When the control command Command is in the "inspection" mode, the speed tolerance input to port _P5 during the "setting" mode is transferred to a register in the microcomputer MC. The pulse MCLK from the rotary encoder RE is captured in the same way as in the "reference" mode. The absolute value of the difference between the traveling speed information at the time of taking in the test value data and that at the time of taking in the reference value data is determined. If the absolute value of this difference is greater than or equal to a predetermined value considering the speed tolerance, the reference value memory rewrite signal RWR
output to BUS. Return to reading the control command Command. Therefore, according to this embodiment, the speed information at the time of taking in the inspection value data of the running printed material to be inspected is compared with the speed information at the time of taking in the reference value data stored in advance, and the difference is determined to be a predetermined value. When the standard value data is exceeded, the standard value data is rewritten.
Misjudgments in inspecting running printed matter due to differences in running speed can be prevented. Next, another embodiment of the present invention using the pixel control data 15 will be described. In rotary printing presses and the like to which the present invention is applied, the width of the printed matter P is not always constant, and printing is often performed by switching to paper of various widths, and in such cases, the width of the inspection area may be changed. It is desirable to change it. For example, as shown in FIG. 32, when the width of printed matter P is switched from A to B, it is desirable to mask portion C. In addition, in such an inspection method, the 33rd
As shown in the figure, there is a logarithmic relationship between the density of the printed pattern and the amount of detected light, so if the judgment level when comparing the reference data and the detected data is set to a fixed level as shown in Figure 33, it is possible to In some areas, the absolute difference between the detected light amount and the judgment level is insufficient,
It becomes impossible to judge whether the pattern is good or bad. Therefore, as shown in FIG. 34, for example, it is desirable to multiply the reference data by a certain ratio (0 to 1) and change it as a function, such as setting the determination level to 10% of the reference data value. Furthermore, at this time, as shown in FIG.
You can mask any part B of the picture A for the screen as a unit, or as shown in Fig. 36,
Taking the pattern C from one printing as a unit, we examine each of the areas that do not require inspection, such as the margin area D, the area E for which inspection at the normal judgment level is sufficient, and the area F that requires particularly strict inspection. It would be even more convenient if the determination level could be changed accordingly and the determination level could be managed in detail in accordance with the picture content of the printed matter. Therefore, in the embodiment described below, masking and judgment level changes can be performed using the same method, and masking can be performed by selecting any part within a pattern. In order to enable partially different judgment levels to be set arbitrarily, the configuration is simple, and inspection is possible while maintaining sufficiently high pattern quality without reducing the yield of finished printed matter. A pixel control data memory having an address corresponding to the data memory is provided, and the determination level of each pixel when an inspection operation is performed by comparing inspection data and reference data is determined based on the data read from this pixel control data memory. It is set in units. The pixel control data memory 15 has the same storage contents as the reference data memory 14 as shown in FIG. 37, similar to the reference data memory 14 explained in FIG. The address a' of the reference data memory 14 is
It is configured such that control data corresponding to the data can be written and read. FIG. 38 shows an example of the feature extraction comparison and determination circuit in this embodiment, in which the first feature extraction comparison and determination circuit 9
It was constructed in . When the printing operation starts and the inspection operation starts, the inspection data ID, the reference data memory 14, and the respective data SD and CD1 from the pixel control data memory 15 are sequentially read out at the same address and sent to each latch 220.
~223 is taken in. Therefore, among these, the positive state reference data SD inputted to the latch 222 and the negative state detection data ID inputted to the latch 223 are sent to the adder circuit 2.
29, an inverter 230, an exclusive OR gate 231, and an adder circuit 232 process the data to become data D _D representing the absolute value of the difference between the test data ID and the reference data SD, which is written into the latch 225. That is, D _D =|ID−SD| is extracted. This data D _D is then supplied from the latch 225 to one input B of the comparator 223. In addition, the reference data taken out from the latch 221
SD is input to the shift register 227, where it is shifted by a predetermined number of bits to a predetermined coefficient K.
(less than 1) becomes the multiplied data KSD and is supplied to one input B of the selection circuit 228. On the other hand, the control data taken into the latch 220
CD1 is read out as is and supplied to the other input B of the selection circuit 228, and the OR gate 22
6 to the selection input S of the selection circuit 228. The selection circuit 228 functions to output the data supplied to the input B when the selection input S is at the "H" level, and outputs the data at the input A when the selection input S is at the "L" level. The comparator 223 compares the data JL of input A and the data D _D of input B, and the data of input B is the data of input A.
It generates an output J that becomes "H" level only when the data is larger than the data. That is, a comparison operation is performed using the data JL as the determination level, and the output J is obtained. Therefore, the computer 21 monitors the output J of the comparator 233, and when it is at "L" level,
At that time, it is determined that there is no defect in the pattern part of the printed matter corresponding to the address of each memory 14, 15, and when it becomes "H" level, it is determined that there is a defect in that part and a predetermined operation is performed. Let's do it. Therefore, the control data currently written to a certain address in the pixel control data memory 15 is (0,
0) _H , that is, all 8 bits are 0. Then, when the inspection data of the picture portion corresponding to this address is read out and inspection is to be performed, the output of the OR gate 226 becomes "L".
Since the level is reached, the selection circuit 228 writes the data KSD of the input A to the latch 224. As a result, the data JL representing the determination level by the comparator 233 at this time is the data KSD, which is obtained by shifting the reference data SD by the shift register 227 and multiplying it by a certain coefficient K.
As a result, the quality of the pattern is determined. That is, at this time, the pattern is inspected using the method described in FIG. 34 in which the judgment level changes as a function of the reference data. Therefore, by writing data (0, 0) _H to an arbitrary address in the pixel control data memory 15, it is possible to obtain an inspection in which the judgment level changes as a function of the reference data in the picture area corresponding to that address. become. In addition, the address specification at this time is
It goes without saying that the process is not limited to each pixel unit, but may be performed in multiple pixel units. Next, suppose that the control data written to a certain address in the pixel control data memory 15 is data other than (0, 0) _H , that is, at least one of the 8 bits is set to 1. Then, when the inspection data of the picture part corresponding to this address is read out, the output of the OR gate 226 becomes "H" level, so at this time, the selection circuit 28 supplies the data CD1 of the input B to the latch 224 as it is. . Therefore, at this time, the comparator 223 operates with the control data CD1 read from that address as the judgment level, and the control data written in the pixel control data memory 15 makes an arbitrary judgment different for each address of the picture. It is possible to have inspections performed at the level. Therefore, for example, as shown in FIG. 36, in a part D of a pattern C of a printed matter to be inspected, which hardly requires inspection, such as a margin, the control data to be written to that address should be made sufficiently large. , thereby increasing the judgment level for this part, ignoring the presence or absence of some defects, and writing relatively small control data to the address corresponding to the important part E, reducing the judgment level. A very strict check is performed on the important portion E, and the determination level is further reduced on the extremely important portion F, thereby enabling fine-grained control over the print finish. At this time, part B of Fig. 35 and part B of Fig. 36
Pixel control data memory 15 corresponding to part D in the figure
If (F, F) _H is written as control data to the address of , the output J of the comparator 233 will always be "H" in this part regardless of the data D _D , so it is possible to detect defects in the pattern. The same result as when masking is not performed is obtained, and the masking operation can be performed on any part of the picture. Next, a method for writing control data into the pixel control data memory 15 will be described. In this embodiment, as is clear from FIG. 2, all operations are controlled and managed by the computer 21, and the computer 21 also writes control data to the pixel control data memory 15. It's becoming like that. Therefore, this write operation will be specifically explained below. When setting an arbitrary value as a determination level in the pixel control data memory 15. As shown in FIG. 39, judgment level data of arbitrary values are sequentially written and set in the pixel control data memory 15 from the computer 21 via the computer interface 17. When setting the reference data multiplied by a certain ratio K (K<1) as the determination level in the pixel control data memory 15. As shown in FIG. 40, the reference data memory 1
4 contents to computer interface 17
is imported into the CPU of computer 21 via
Thereafter, each pixel of the reference data is multiplied by a certain ratio K (K<1) within the CPU to process it into data representing the determination level. In this case, in areas where the density level is low (where the digital value is small), when multiplied by the ratio K, it may become 0, so it is necessary to determine the lowest value that can be set, or to use a standard value. Ratio K to data
It is a good idea to add a certain value after multiplying by . The thus processed data representing the determination level for each pixel is set in the pixel control data memory 15 from the CPU via the computer interface 17. Note that it goes without saying that a hardware configuration may also be used. When monitoring the reference data as an image, positioning an arbitrary part of the pattern using a cursor, etc., and setting a determination level for each arbitrary part in the pixel control data memory 15. As shown in FIG. 40, the reference data memory 1
4 contents to computer interface 17
The data is imported into the CPU via the , and then input to the CRT so that it can be monitored as an image. and,
Move the cursor to specify a predetermined part of the reproduced image, import that position to the CPU,
An arbitrary determination level at that position is set in the pixel control data memory 15 from the CPU via the computer interface 17. When extracting the outline of a picture from reference data, setting a special judgment level for this part, and setting it in the control data memory. As shown in FIG. 40, the reference data memory 1
4 contents to computer interface 17
The data is input to the CPU via the CPU, and digital image processing calculations are performed to extract the outline of the pattern from the data. In this way, an arbitrary judgment level is determined for the address of the extracted contour part,
Data is set in the pixel control data memory 15 via the computer interface 17. As a method for extracting components representing vertical or horizontal contours from the pattern of image data, for example, there is a method using a spatial filter technique that is commonly used in digital image processing technology. Accordingly, you can easily reach your goal. Now, centering on any point fij on the pattern, 3x
If we use the density of the intermediate point in the square area of 3,
The spatial filter used for vertical contour extraction is

[Formula] or

[Formula], and the spatial filter used for horizontal contour extraction is:

[Formula] or

[Formula] becomes. Also, if you simply want to find the outline of a picture, the spatial filter can be used to calculate differences in four directions using the Laplacian method.

[Formula], and in the method of taking the difference in 8 directions,

[Formula]. The reason for extracting the outline of the picture and setting a special determination level as shown above is as follows. In other words, in parts of the pattern where the density changes rapidly, the detection operation is greatly affected by the positional deviation that occurs during defect detection, and as a result of the positional deviation, even though there is no defect, the detection operation is greatly affected. There is a greater possibility that a detection operation will be performed assuming that there is a defect, making it more likely that a malfunction will occur. Therefore, by detecting the outline of the picture and setting the judgment level for this part higher than for other parts, it is possible to reduce malfunctions due to positional deviation without increasing the overall judgment level too much. This is because it can be done. In addition, in such an inspection, in order to make it easier to detect defects in the rotational direction such as doctor streaks that occur during the gravure printing stage, the size of the detection area of one pixel is reduced horizontally, as shown in Figures 41A and B. It is advantageous to be longer in the vertical direction than in the vertical direction, and in this case, even if the position is shifted by the same distance (△X, △Y), the effect on the detection area of one pixel photoelectric element is as shown in A of Fig. 42. As is clear from and B, the deviation in the horizontal direction X is larger than the deviation in the vertical direction. Therefore, as mentioned above, by detecting the contour part and setting the horizontal judgment level to be higher than the vertical one in that part, malfunctions due to positional deviations will be reduced, and the overall defect detection accuracy will be improved. Even if it is not lowered, the risk of increased malfunction can be eliminated. Note that the control data setting operations explained above do not need to be performed individually, but should be performed in conjunction with hardware and software setting methods in order to reduce data setting time and reduce the burden on hardware manufacturing. You can do it like this. An example of how to set data in such a case is shown in Part 4.
This is shown in the flowchart in Figure 3. Although not specifically explained above, the buffer memory 16 in the block diagram of FIG. interface 19
is for outputting internal digital data to the monitor 20. The computer 21 is, for example, what is called a personal computer, and it goes without saying that it is provided to control the operation of the entire inspection apparatus. In addition, in this embodiment, the inspection position is printed on the cylinder CY.
As mentioned above, it goes without saying that this can be applied to offline inspection and inspection of sheet-fed printed matter as long as it is equipped with a mechanism that can grasp the absolute position of the pattern. As explained above, according to the present invention, it is possible to provide an apparatus that can inspect printed matter in real time while maintaining extremely high accuracy even when the printed matter is running at a fairly high speed, such as with a rotary printing machine. can.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram showing the principle of an example of a printed matter inspection method using a digitization method, FIG. 2 is a conceptual diagram of a reference data memory, and FIG. 3 is an illustration of an embodiment of a printed matter inspection apparatus according to the present invention. A block diagram showing the entire system, FIG. 4 is a circuit diagram showing one embodiment of the traveling position signal input circuit, FIG. 5 is a time chart for explaining the operation, and FIG. 6 is a circuit diagram showing one embodiment of the scanning direction signal input circuit. Fig. 7 is an explanatory diagram showing the relationship between the inspection surface of the printing cylinder and each signal; Fig. 8 is a circuit diagram shown in Fig. 7;
FIG. 9 is an explanatory diagram showing address division on the inspection surface, FIG. 9 is a circuit diagram showing an example of an address generation circuit, FIG. 10 is an explanatory diagram showing an example of characteristics required for an analog-to-digital converter, and FIG. 12 is a circuit diagram showing an embodiment of a digital input interface, FIG. 12 is a circuit diagram showing an embodiment of an image enhancement circuit,
Fig. 13 is an explanatory diagram showing the configuration and signals of the data input system, Fig. 14 is a circuit diagram showing an example of common parts, and Fig. 15 is an explanatory diagram showing comparison and judgment operations in pixel units. , FIG. 16 is a circuit diagram showing an embodiment of the first feature extraction comparison and judgment circuit, and FIG. 17 is an explanatory diagram showing a comparison and judgment operation based on the sum of pixels.
Fig. 18 is a circuit diagram showing an embodiment of the second feature extraction comparison/judgment circuit, Fig. 19 is an explanatory diagram showing a comparison/judgment operation based on the summation of pixels in a specific direction, and Fig. 20 is a circuit diagram showing an example of the second feature extraction comparison/judgment circuit. FIG. 21 is a circuit diagram showing an embodiment of the comparison judgment circuit. FIG. 21 is a circuit diagram showing an embodiment of the comprehensive judgment circuit. FIG. An explanatory diagram showing changes in pixel light intensity, Fig. 24 is an explanatory diagram showing the operation of the picture position detector, Figs. 25 and 26 are explanatory diagrams of position detection marks, and Fig. 27 is a reference data memory rewrite signal. A circuit diagram showing an embodiment of the generating circuit, FIG. 28 is a flowchart for explaining its operation, FIG. 29 is an explanatory diagram showing the relationship between the running speed of printed matter and the actual scanning direction, and FIG. 30 is a flowchart for explaining the operation of the generating circuit. An explanatory diagram showing a change in the reading state of a pattern due to a change in traveling speed, FIG. 31 is a flowchart showing an example of the reference data memory rewriting operation depending on the traveling speed, and FIG. 32 shows the necessity of masking in printed matter inspection. 33 and 34 are characteristic diagrams illustrating the relationship between pattern density and light amount and the determination level, FIGS. 35 and 36 are explanatory diagrams showing the effects of this embodiment, and FIG. A conceptual diagram of the pixel control data memory, FIG. 38 is a circuit diagram showing another embodiment of the first feature extraction comparison and determination circuit, and FIGS. 39 and 40 are diagrams of data setting operations for the pixel control data memory, respectively. FIGS. 41A and 41B are explanatory diagrams showing the shape of one pixel in the detection area, FIGS. 42A and B are explanatory diagrams showing the difference in positional deviation detection sensitivity depending on the direction, and FIG. 43 is an explanatory diagram showing the embodiment. 7 is a flowchart showing an example of data setting operation. P...Printed material, CY...Printing cylinder, RE...
Rotary encoder, 1... Image sensor camera (IS), 3... Picture position detector.

Claims (1)

[Scope of Claims] 1. A memory capable of storing image data read from the pattern of the running printed material at a predetermined time, and the image data read from the memory is read from the reference data and the remaining pattern of the running printed material. In a printed matter inspection device that uses image data as inspection data and determines the quality of printing based on a comparison between these reference data and inspection data, the characteristics of the reference data and the characteristics of the inspection data are determined for each corresponding pixel. A first feature extraction/comparison/judgment means for extracting and comparing to determine the quality of printing, and printing by extracting and comparing the features of the reference data and the features of the inspection data as the sum of pixels in a predetermined range within the same picture. a second feature extraction comparison/judgment means for determining the quality of the reference data and the features of the inspection data as the sum total of pixels arranged on the same straight line along the running direction of the printed matter within the same picture; a third feature extraction comparison and determination means for determining the quality of printing by
A comprehensive judgment means is provided for comprehensively judging the judgment results of each of the feature extraction and comparison judgment means to make a final judgment on the quality of the printed matter, and the quality of the printed matter is judged based on the judgment result of the comprehensive judgment means. ,
a speed detection means for detecting the running speed of the running printed matter; a means for retaining the speed data detected by the speed detecting means; A calculation means for calculating the difference from the speed data, a comparison means for comparing the result of the calculation means with a predetermined value, and a comparison means for comparing the inspection data read from the pattern of the running printed matter at this time based on the result of the comparison means. A printed matter inspection device characterized in that it is provided with means for newly rewriting reference data. 2. In claim 1, there is provided means for detecting a pattern position in a direction perpendicular to the traveling direction of the traveling printed matter, and means for retaining the position data detected by the means, and the reference data is stored in the memory. The above picture position at the time of writing,
An apparatus for inspecting printed matter, characterized in that, when a difference between the inspection data and the pattern position exceeds a predetermined value when reading the inspection data, reference data is rewritten to the memory. 3. In claim 1 or 2, a pixel control data memory having an address corresponding to each pixel unit or a plurality of pixel units of image data, and determining a tolerance difference between the reference data and the detected data. data setting means for writing control data into the pixel control data memory in units of each pixel or a plurality of pixels of image data; An inspection apparatus for printed matter, characterized in that a determination level in at least one of the operations for comparing data and detected data is set based on pixel control data read from the pixel control data memory. 4. In claim 3, it is provided that: means for identifying pixel control data read from the pixel control data memory; arithmetic means for performing predetermined arithmetic processing on the reference data read from the memory; and the reference data and the detected data. A selection means is provided for selecting a judgment level in the comparison operation by switching to either one of the output data of the arithmetic means and the pixel control data read from the pixel control data memory, and according to the identification result by the identification means. 1. An inspection apparatus for printed matter, characterized in that said selection means is controlled by said selection means. 5 In claim 3 or 4,
A level determining means is provided for determining the level of the pixel control data read from the pixel control data memory, and when the level of the data exceeds a predetermined value, the determination level in the comparison operation between the reference data and the detected data is set to a sufficient level. A printed matter inspection device characterized in that the printed matter is configured to be set at a high level. 6. In any one of claims 3 to 5, the data setting means comprises a monitor means for reproducing reference data read from the memory on a predetermined image projection surface, and an image projection surface of the monitor means. and means for specifying any part of the pixel control data memory to write data to the corresponding address of the pixel control data memory. 7. In any one of claims 3 to 5, the data setting means extracts the outline of the image based on the reference data read from the reference data memory and sets the image to the corresponding address in the pixel control data memory. 1. An inspection device for printed matter, characterized in that it is comprised of means that allows data to be written to.