JPH0147822B2 - - 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. Therefore, instead of the conventional visual inspection method, a method has been proposed and put into practical use that uses image processing technology to automatically inspect printed matter by comparing image patterns. Figure 1 shows the principle of an example of such an inspection method by comparing image patterns, in which 1 is a printed matter, 2 is a rotating drum, 3 is an image sensor camera, and 4 is an analog/digital (A/D). Converter, 5 is a changeover switch, 6 is a reference data memory,
7 is a detection data memory. The printed matter 1 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 rotating drum 2. An image sensor camera (hereinafter referred to as an IS camera) 3 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 traveling direction Y of the printed material 1. Scan and extract image signals. The one-dimensional image signal read from the printed matter 1 is digitized for each predetermined pixel by an A/D converter 4 and supplied to a switch 5. The reference data memory 6 and the detection data memory 7 have storage contents as shown in FIG.
The structure is such that digitized density information can be written and read out for each pixel. Therefore, if the switch 5 is now switched to the reference data memory 6 as shown in the figure, and a predetermined pattern part of the printed matter 1 is imaged by the IS camera 3, 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 6, which are sequentially repeated by the movement of the printed matter 1 in the running direction Y, and the density is written to the addresses divided in the running direction of the reference data memory 6. The image data is written as information, and image data of a predetermined range of the image printed on the printed material 1 is eventually written into the reference data memory 6. Moreover, if the switch 5 is switched to the lower side opposite to that shown in the figure, image data of a predetermined range of the pattern of the printed matter 1 will be written in the same way. 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 5 is switched to the reference data memory 6, and at that time, the data obtained from the printed matter 1 is transferred. The detected image data is written to the reference data memory 6 as reference data, and from then on, the switch 5 is used as the detection data memory 7.
If the image data is switched to , image data sequentially read from the printed material 1 is input to the memory 7 one after another, and new image data is sequentially rewritten as detection data. Therefore, after that, the density information of the corresponding addresses in the reference data memory 6 and the detection data memory 7 are sequentially read out and compared, and it is determined whether or not they match. The quality of printing can be continuously determined, and the inspection can be performed with a level of reliability that is incomparably higher than that of visual inspection. By the way, in printing machines to which such an inspection method is applied, the width of the printed matter 1 is not always the same, but it is often printed by switching to sheets of different paper widths, and in such cases, the inspection area It is desirable to vary the width of. for example,
As shown in FIG. 3, when the width of the printed product 1 is switched from A to B, masking of portion C is desirable. In addition, in such an inspection method, since there is a logarithmic relationship between the density of the printed pattern and the amount of detected light as shown in Figure 4, the judgment level when comparing the standard data and the detected data is shown in Figure 4. If a fixed level is used as shown in , the absolute difference between the detected light amount and the determination level will be insufficient in the low density region, making it impossible to determine the quality of the image. Therefore, as shown in FIG. 5, for example, it is desirable to set the determination level to 10% of the reference data value, or to multiply the reference data by a certain ratio (φ~1) and change it as a function thereof. Therefore, for example, regarding masking technology, there is
JP-A No. 54-97484 has proposed a technique for varying the determination level. However, in these methods, for example, the sixth
As shown in the figure, you can mask any part B of the picture A for one screen, or
As shown in the figure, a pattern C from one print is taken as a unit, and there are areas that do not require inspection, such as margins D, areas E that are sufficient for inspection at the normal judgment level, and areas that require particularly strict inspection. In addition, it is not possible to change the judgment level according to each part F, and to manage the judgment level in detail according to the image content of the printed matter, and masking and changing the judgment level are performed in completely different ways. Therefore, the disadvantage was that the structure became complicated. An object of the present invention is to eliminate the drawbacks of the prior art described above, to enable masking and change of judgment level to be performed by the same means, and to enable masking by selecting any part within a picture. It is possible to arbitrarily set partially different judgment levels depending on the top pattern, which allows for a simple configuration and maintains sufficiently high pattern quality without reducing the yield of finished printed matter. To provide a method for inspecting printed matter. In order to achieve this object, the present invention provides a control data memory having an address corresponding to the reference data memory, and the determination level when an inspection operation is performed by comparing detected data and reference data is set from the control data memory. It is characterized in that it is set for each pixel or for a plurality of pixels based on the read data. Embodiments of the printed matter inspection method according to the present invention will be described below with reference to the drawings. FIG. 8 shows an embodiment in which the inspection method according to the present invention is realized by computer control, and the rotating drum 2, IS camera 3, A/D converter 4, reference data memory 6, detection data memory 7, etc. This is the same as the conventional example. In FIG. 8, 9 is a rotary encoder for detecting the rotational phase of the rotary drum 2, 10 is a running position signal input circuit, 11 is a scanning direction signal input circuit, 12 is an address generation circuit, and 13
1 is a computer interface, 14 is an arithmetic unit, 15 is a computer, and 16 is a control data memory. A rotary encoder 9 generates a signal representing the rotational angular position of the rotary drum 2. Therefore, by checking the signal from the rotary encoder 9, the traveling position of the printed material can be determined, and image data from a predetermined portion of the pattern can be captured. The signal from this rotary encoder 9 is transmitted to the traveling position signal input circuit 1.
0 to the internal bus BUS. The scanning direction signal input circuit 11 functions to send a signal representing the scanning direction by the IS camera 3 to the internal bus BUS. The address generation circuit 12 includes a reference data memory 6, a detection data memory 7, and a control data memory 1.
The address signals necessary for the operation of 6 are transferred to the internal bus BUS.
It functions to send out to. The arithmetic unit 14 compares the detection data read out from the detection data memory 7 with the reference data from the reference data memory 6 to determine the quality. The computer 15 is called a personal computer, for example, and consists of a central processing unit CPU, a memory MEM, a keyboard KB, a display CRT, etc. It is connected to the internal bus BUS via the computer interface 13 and follows a predetermined program. Then various processing is performed. The control data memory 16 has the same storage contents as the reference data memory 6 as shown in FIG. 9, similar to the detection data memory 7 described in FIG. 2, and is similarly divided into the printing width direction and the running direction. address
It is configured such that control data corresponding to address a of the reference data memory 6 can be written to and read from a'. FIG. 10 shows an embodiment of the arithmetic unit 14, with two
0 to 25 are 8-bit data latch circuits, 26 is an OR gate array configured to allow 8-bit input, 27 is a shift register, 28 is a selection circuit, 29 and 32 are adder circuits, and 30 is also an inverter, 31 is also an exclusive OR gate, and 33 is a comparator. Next, the operation of this embodiment will be explained. When the printing operation starts and the inspection operation begins, each data from the detection data memory 7, the reference data memory 6, and the control data memory 16 is sequentially read out at the same address by the program of the computer 15 and sent to each latch 20. ~23. Therefore, among these, the positive state reference data D _S input to the latch 22 and the negative state detection data D _I input to the latch 23 are input to the arithmetic circuit 29.
Inverter 30, exclusive or gate 3
1. It is processed by the circuit consisting of the adder circuit 32 to become data D _D representing the absolute value of the difference between the detection data D _I and the reference data D _S , and is written into the latch 25. That is, D _D =|D _I −D _S | is extracted. This data D _D is then supplied from the latch 25 to one input B of the comparator 33. In addition, the reference data taken out from the latch 21
D _S is input to the shift register 27, where it is shifted by a predetermined number of bits to a predetermined coefficient K(1
(smaller) becomes the multiplied data KD _S and is supplied to one input B of the selection circuit 28. On the other hand, the control data D _C taken into the latch 20
is read out as is and supplied to the other input A of the selection circuit 28, and is also supplied to the selection input S of the selection circuit 28 via the OR gate 26. The selection circuit 28 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 33 compares the data JL of input A with the data D _D of input B, and generates an output J that becomes "H" level only when the data of input B is greater than the data of input A. That is, a comparison operation is performed using the data JL as the determination level, and the output J is obtained. Therefore, the computer 15 monitors the output J of the comparator 33, and when it is at the "L" level, it determines that there is no defect in the image part of the printed material corresponding to the address of each memory 6, 7, 16 at that time. However, when it reaches the "H" level, it is determined that there is a defect in that part and a predetermined operation is performed. Therefore, the control data currently written to a certain address in the control data memory 16 is (0, 0).
Assume that _H , that is, all 8 bits are 0. Then, when the detection data of the picture part corresponding to this address is read out and an inspection is to be performed, the output of the OR gate 26 becomes "L" level, so the selection circuit 28 selects the data of the input A.
Write KD _S to latch 24. As a result, the data JL representing the determination level by the comparator 33 at this time is data KD _S , which is obtained by shifting the reference data D _S by the shift register 27 and multiplying it by a certain coefficient K. The quality of the pattern will be judged. That is, at this time, the pattern is inspected using the method described in FIG. 5 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 control data memory 16, an inspection method can be obtained 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, assume that the control data written to a certain address in the control data memory 16 is data other than (0, 0) _H , that is, at least one of the 8 bits is set to 1. Then, when the detection data of the picture part corresponding to this address is read out, the output of the OR gate 26 becomes "H" level, so at this time, the selection circuit 28 supplies the data _DC of the input B to the latch 24 as is. do. Therefore, at this time, the comparator 33 operates with the control data D _C read from that address as the judgment level, and the control data written in the control data memory 16 sets an arbitrary judgment level that differs for each address of the picture. inspection can be carried out. Therefore, for example, as shown in FIG. 7, 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 at that address should be made sufficiently large. ,
As a result, the judgment level for this part is set high, and the presence or absence of some defects is ignored, and relatively small control data is written to the address corresponding to the important part E, and the judgment level is made small. A very strict check is performed on the most important portions E, and the determination level is further reduced on the extremely important portions F, thereby enabling fine-grained control over the printing finish. At this time, if (F, F) _H is written as control data to the address of the control data memory 16 corresponding to part B in FIG. 6 or part D in FIG. Since the output J of 33 is always “H” regardless of the data D _D ,
The same result as when no pattern defect detection is performed is obtained, and the masking operation can be performed on any part of the pattern. Next, a method for writing control data into the control data memory 16 will be explained. In this embodiment, as is clear from FIG. 8, all operation control and operation management are performed by the computer 15, and writing of control data into the control data memory 16 is also performed by the computer 15. It's getting old. Therefore, this write operation will be specifically explained below: When setting an arbitrary value to the control data memory 16 as a determination level. As shown in FIG. 11, judgment level data of arbitrary values are sequentially written and set in the control data memory 16 from the CPU via the computer interface 13. The control data memory 1 uses the reference data multiplied by a certain ratio K (K<1) as the determination level.
When setting to 6. As shown in FIG.
The content of is read into the CPU via the computer interface 13, and then a certain ratio K (K<1) is calculated for each pixel of the reference data inside the CPU.
is processed into data representing the judgment level. At this time, in areas where the density level is low (where the digital value is small), the ratio K
Since there is a case where the value becomes 0 when multiplied by the ratio K, it is better to set the lowest value that can be set or to multiply the reference data by the ratio K and then add a constant value. Data representing the judgment level for each pixel processed in this way is sent from the CPU to the control data memory 16 via the computer interface 13.
Set to . Further, as in this embodiment, in the arithmetic unit 14 in FIG.
Figure) Needless to say, it can be configured. When monitoring reference data as an image, positioning an arbitrary part of a pattern using a cursor, etc., and setting a determination level for each arbitrary part in the control data memory 16. As shown in FIG.
The contents are taken into the CPU via the computer interface 13, and input into the CRT so that they can be monitored as images. Then, move the cursor to specify a predetermined part of the reproduced picture pattern, import that position into the CPU, and set an arbitrary judgment level at that position from the CPU to the control data memory 16 via the computer interface 13. Ru. When extracting the outline of a pattern from reference data, setting a special judgment level for this part, and setting it in the control data memory. As shown in FIG.
The contents of the image are taken into the CPU via the computer interface 13, and digital image processing calculations are performed to extract the outline of the picture from the data. An arbitrary determination level is determined for the address of the extracted outline, and the data is set in the control data memory 16 via the computer interface 13. From the pattern of such image data, its vertical direction,
Alternatively, as a method for extracting a component representing a horizontal contour, for example, there is a method using a spatial filter technique that is generally used in digital image processing technology, and the purpose can be easily achieved using this method. Now, if we use the density of each point in a 3x3 square area centered on an arbitrary point f _ij of the pattern, the spatial filter used to extract the contour in the vertical direction is:

[Formula] or

[Formula], and the spatial filter used to extract the horizontal contour is:

[Formula] or

[Formula] becomes. Also, if you simply want to find the outline of a pattern, 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, defects may be detected even though no defects exist. There is a greater possibility that the detection operation will be performed assuming that the temperature is hot, 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 method, 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 increased as shown in Fig. 13A and B. It is advantageous to make it longer in the vertical direction than in the horizontal direction. 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 shown in Figure 14. As is clear from A and B, the deviation in the horizontal direction X is larger than the deviation in the vertical direction Y. 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 a setting method in such a case is shown in a flowchart in FIG. Although this embodiment has been described as a method in which the detected data is once written in the memory and then read out and compared, the present invention is implemented as a real-time processing method in which the detected data is immediately compared with the reference data without using the memory 7. Needless to say, it's good. Further, this also applies to the conventional example shown in FIG. As explained above, according to the present invention, it is possible to selectively designate and mask any part of the pattern of a printed matter to be inspected, and to perform inspection at different arbitrary judgment levels. It is possible to provide a printed matter inspection method that eliminates the shortcomings of the prior art, maintains sufficiently high defect detection accuracy for necessary parts of a pattern, and does not reduce the overall yield.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram showing an example of a printed matter inspection method equipped with a reference data memory, FIG. 2 is an explanatory diagram of the memory, FIG. 3 is an explanatory diagram showing the necessity of masking in printed matter inspection, and FIGS. Fig. 5 is a characteristic diagram illustrating the relationship between pattern density and light amount and the judgment level, Figs. 6 and 7 are explanatory diagrams showing the effects of the present invention, and Fig. 8 is a computer-based method for inspecting printed matter according to the present invention. FIG. 9 is an explanatory diagram of the control data memory, FIG. 10 is a block diagram showing an embodiment of the arithmetic unit, and FIGS. An explanatory diagram showing one example of the data setting operation for the control data memory, FIGS. 13A and B are explanatory diagrams showing the shape of one pixel in the detection area, and FIGS. 14A and B are depending on the direction of positional deviation detection sensitivity. An explanatory diagram showing the difference, FIG. 15 is a flowchart showing an example of control data setting operation. DESCRIPTION OF SYMBOLS 1... Printed matter, 2... Rotating drum, 3... Image sensor camera, 4... Analog-digital converter, 6... Reference data memory, 7... Inspection data memory, 9... Rotary encoder, 10...
...Traveling position signal input circuit, 11...Scanning direction signal input circuit, 12...Address generation circuit, 13...
Computer interface, 14... Arithmetic unit, 15... Computer, 16... Control data memory.

Claims (1)

[Claims] 1. A reference data memory in which image data read from a reference printed matter is written as reference data, and image data read from the printed material to be inspected is used as detection data and read from the memory. A control data memory having an address corresponding to each pixel unit or a plurality of pixel units of image data in a printed matter inspection method that determines the quality of printing by comparing each pixel unit or a plurality of pixel units with reference data. control data for determining the tolerance of the reference data with respect to the detected data is written in a predetermined address of the control data memory, and the reference data and the detection data are determined according to the identification result of the control data read from the control data memory. The judgment level in the detection data comparison operation is
The control data read from the control data memory and the output data obtained by performing predetermined arithmetic processing on the reference data read from the reference data memory are switched to one of the control data and the output data obtained by performing predetermined arithmetic processing on the reference data read from the reference data memory. A method for inspecting printed matter, characterized in that the method is configured to perform a judgment based on a judgment level. 2. In claim 1, when the level of the control data read from the control data memory is determined and the level of the control data is equal to or higher than a predetermined value, the comparison operation between the reference data and the detected data is performed. A printed matter inspection method characterized in that a determination level is set to a sufficiently large level. 3 In claim 1 or 2,
Writing of the control data to a predetermined address of the control data memory involves inputting the data read from the reference data memory to a monitor having a predetermined image projection surface, and specifying any part of the image projection surface of this monitor. A method for inspecting printed matter, characterized in that the method is configured such that data is written to a corresponding address of the control data memory. 4 In claim 1 or 2,
A method for inspecting printed matter, characterized in that the data written in the control data memory is data obtained by extracting the outline of an image based on the reference data read from the reference data memory.