Method for automated alignment and register measurement using circular measuring marks
11267264 · 2022-03-08
Assignee
Inventors
- Jan Krieger (Heidelberg, DE)
- Immanuel Fergen (Karlsruhe, DE)
- Christoph Godau (Mannheim, DE)
- Timo Volk (Mannheim, DE)
- Manfred Schneider (Bad Rappenau, DE)
Cpc classification
B41J11/42
PERFORMING OPERATIONS; TRANSPORTING
International classification
Abstract
A method for automated alignment and register measurement in a printing press provides for test patterns having multiple color separations to be printed by the printing press on a printing substrate, recorded by using at least one image sensor of an image acquisition system as a digital overall image, evaluated by a computer with respect to an alignment/register offset and then corrected by the computer for the alignment/register offset. Circular measuring marks having known diameter for each color separation are integrated into the test patterns and the computer ascertains the center position of each circular measuring mark with subpixel accuracy and thus computes the alignment/register offset by cutting out an image region having at least one circular measuring mark from the digital overall image and determining parameters of a model of a printing point of the circular measuring mark from the digital overall image.
Claims
1. A method for automated alignment and register measurement in a printing press, the method comprising the following steps: using the printing press to print test patterns having multiple color separations on a printing substrate; using at least one camera of an image acquisition system to record the print test patterns as a digital overall image; using the computer to evaluate the digital overall image with respect to an alignment/register offset and then using the computer to correct the alignment/register offset; integrating circular measuring marks of known diameter into the test patterns for each color separation; and using the computer to ascertain a center position of each circular measuring mark with subpixel accuracy for computing the alignment/register offset by cutting out an image region having at least one circular measuring mark from the digital overall image and determining parameters of a model of a printing point of the circular measuring mark from the digital overall image.
2. The method according to claim 1, which further comprises placing the circular measuring marks on the printing substrate so as to be completely acquired by a single camera and imaged in a single digital overall image.
3. The method according to claim 1, which further comprises using closed circular disks or open circular rings of known diameter as the circular measuring marks for each color separation.
4. The method according to claim 1, which further comprises carrying out the method by: ascertaining an alignment offset between printing bars having printheads disposed adjacent one another in an inkjet printing press; and placing the circular measuring marks in lines horizontally or vertically on the printing substrate.
5. The method according to claim 4, which further comprises using the computer for ascertaining the alignment offset between the printing bars by: determining deviations of the center position of the circular measuring mark from known ideal positions; applying an outlier-robust regression method to average the deviations over the printing substrate; and using the computer to ascertain and compensate for the alignment offset by oppositely driving the printing bars.
6. The method according to claim 5, which further comprises not printing the circular measuring marks in an overlap region between two printheads.
7. The method according to claim 4, which further comprises not printing the circular measuring marks in an overlap region between two printheads.
8. The method according to claim 1, which further comprises carrying out the method for ascertaining the alignment/register offset in an offset printing press by: integrating the circular measuring marks into existing printing control strips; and using the circular measuring marks to replace previous color measuring fields.
Description
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING
(1)
(2)
(3)
(4)
(5)
DETAILED DESCRIPTION OF THE INVENTION
(6) Referring now in detail to the figures of the drawings, in which elements corresponding to one another are provided with the same reference signs, and first, particularly, to
(7)
(8) Lines of filled circular disks 9 are disposed on a printed sheet, so that each printhead produces circular disks 9. At least three circular disks 9 are distributed per printhead in this case, in such a way that they are not printed in the stitching region. Each printhead is to print at least two complete sets of circular disks 9 located adjacent one another. The circular disks 9 are to be distributed in this case over the sheet in such a way that the largest possible region is spanned in the y direction. This guarantees a high resolution in both coordinate directions (x: printing bar direction, y: paper transportation direction). The circular disks 9 have to be sufficiently large that they provide good results for the method according to the invention even upon the presence of white lines or diagonally spraying nozzles. The disk size may be determined experimentally with the aid of a simulation. A target size of 60 camera pixels or, at 670 dpi camera resolution, >2.2 mm diameter, is sufficient for images of high quality from the above-described camera system 5. With some spacing in relation to the large lines and as close as possible to the beginning of the printing pattern, one additional offset point 8 is disposed per camera 5, which enables a referencing (“grid point” 8) for the point search. The requirements are the same as already described.
(9) For the following method, a partial method is necessary, which determines the center point of a circular disk 9 with subpixel accuracy. The method used for this purpose is performed as follows:
(10) Since the printing color for the circular disk 9 to be examined is known, the image can be ideally converted into a high-contrast grayscale image with the aid of this information, for example, by selecting the channel R, G, or B having highest contrast to the printed material. A region (ROI) approximately two times larger than the actual circular disk 9 is cut out of the overall image converted into grayscale. In the ROI having the circular disk 9, an edge detection is carried out, so that the edge of the circular disk 9 remains as a 1-pixel-wide line in a binary image 10. In order to filter out line artifacts in the circular disks 9, which now stand out as double lines, a further filter is applied, which firstly finds vertical lines in the binary image which is formed of one pixel and are at least nine pixels tall and then subtracts them from the mask, which results in an error-free circular disk 9 after the edge detection. The limit of at least nine pixels is selected as configurable for adaptability of the method to other camera resolutions. The binary image is expanded by a cross through the center point of the circular disk 9 approximately determined as the mass center of gravity. The result is referred to as MASK. The MASK is finally widened by dilation to three to five pixels. The determination of how many pixels it is precisely widened by is dependent on the imaging power and/or the step response of the camera system being used. A nonlinear least-squares fit is now carried out on the masked ROI data, wherein the parameters of a model of the printed point are determined. The model is defined by a radial intensity curve f(r) from the center point (x0, y0) of the circular disk outward. The radial intensity curve f(r) is basically defined as a step from the central color value (A0+A1) to the background value A0 at the radius r0, which is also widened to a width w in order to take into consideration the limited imaging power of the objective lens. Moreover, an asymmetry factor a can be taken into consideration in order to take into consideration, for example, unequal resolutions in the x and y directions. This also relates, for example, to a reduction of the resolution in one direction to compensate for higher printing speeds.
(11) Firstly, reasonable starting parameters are selected for all of these parameters. Such parameters are, for example, the mass center of gravity of the image for x.sub.0, y.sub.0, the radius for r.sub.0 expected from the printed image, machine experiential values for w, etc. The fit is then executed by using a standard method of numerics, for example, a Levenberg-Marquardt method.
(12) The results are checked for meaningfulness. Thus, for example, the radius or the center has to be located in the expected region. If this is not the case, the corresponding circular disk 9 is discarded in case of doubt and not used for the further analysis. This algorithm also still functions if the circular disk 9 is not completely in the ROI, but rather only a portion >50-60% is visible. However, the accuracy then possibly suffers. This can be taken into consideration through a weighting or a scoring of the results, however.
(13) This standard method for the evaluation of the test pattern 10 having circular disks 9 is to be used for the in-line measurement and regulation of the alignment/register deviation for a sheet-fed offset printing press 4. Various adaptations of the alignment/register marks previously used are required for this purpose. The calibration points/circular disks 9 are therefore now integrated into the previous color control strips 19.
(14)
(15) One example for the use of a color control strip 19 having integrated circular disks 11 is described in greater detail hereafter:
(16) A horizontal line having circular disks 11 is used, with at least one circle per printing unit and integration into the known printing control strips is possible.
(17) The length of the line is: L=2*xa+4*d+3*xi and the diameter of a circular disk 11 is d=(L−2*xa−3*xi)*¼, with xi=0.8 mm as the spacing between the circular disks 11 and xa=0.5 mm for the spacing of the line in relation to the closest object of the color control strip 19 which then results in a length of the line L=13 mm with d=2.275 mm for the maximum possible diameter d of a circular disk 11. In general, the possible diameter is in the range of 0.2 to 5 mm.
(18) The described method for analyzing the circular disks 11 can then be carried out by using this adapted color control strip 19, and therefore alignment/register deviations can be computed without having to use separate register marks. The integration of the circular disks 11 into the color control strips 19 does not obstruct its actual task in the case of the color measurement or color control, since the circular disks 11 also depict various color separations and are suitable for the color measurement.
(19) The application appears somewhat different for use in inkjet printing presses 4. The alignment between two printing bars 15, which are each formed of printheads disposed adjacent one another, is to be measured in this case. The alignment/register offset between the printheads of one color is thus not meant, but rather between the entire printing bars 15 of different colors. This roughly corresponds to the alignment between the color separations.
(20) The register marks required for this purpose, i.e., adapted inkjet test printing patterns 12 of various color separations, are formed of circular disks 11 having a minimum diameter, so that the presence of defective printing nozzles in the printhead also has no influence on the alignment measurement.
(21) The regulation method for the alignment of the printing bars 15 in relation to one another is performed as follows in this case and is schematically illustrated once again with respect to the results in
(22) 1. Determine alignment deviations of one or each color in relation to a reference color (typically: BLACK). a. Roughly find and cut out all circular disks 11, this can optionally also be carried out through a circle feature detection (for example, a Hough transformation). b. Determine the center position of each circular disk 11 with subpixel accuracy. c. Convert center positions with the aid of coordinate transformation into real coordinates (for example, millimeters or printing pixels). d. For deviations transverse to the orientation of the test strip, a straight line can be laid through the positions found of the circular disks 11. For this purpose, an outlier-robust regression method, such as iteratively-reweighted-least-squares (IRLS), or a normal least-squares fit is used. This is carried out for measuring and reference colors and the spacing of the straight lines can then be computed at any arbitrary point to determine the alignment transverse to the orientation of the test strip. It can possibly be taken into consideration in this case that measuring and reference points were not printed on a line. The two straight lines also enable the angle between the printing bars 15 to be measured. Alternatively, a method as described in e. can also be used. e. For deviations along the test strip, the expected ideal position of the circular disks 11 is determined (for example, by a predetermined calibration of the cameras 5 or a local calibration or coordinate transformation computed from surrounding reference points between the printed image and the camera image). The deviation of the found position from the expected position is then the alignment error at the measuring mark. f. Optionally: averaging of the measurements from various sheets and/or various measuring patterns on one sheet. g. Determine X/Y offset, so that the alignment deviations are minimized over the entire sheet—i.e., the deviations do not set the left side to zero, for example, but the right side is excessively large in exchange, as seen in
(23) 2. Adjust determined X/Y offset in machine controller in the opposite direction to compensate for the measured deviation.
(24) 3. Further measurement: a. Deviation less than limit->FINISHED. b. Deviation greater than limit->back to step 1.
(25) In a further preferred embodiment variant, open circular rings having suitable diameters can moreover also be used for both applications instead of the circular disks 9, 11, if they can also be integrated into already known printing control strips 19 or register marks. They are theoretically less susceptible to printing-technology artifacts.
(26) The advantages of the method according to the invention for both applications, i.e., offset and inkjet, are that a separate measuring device is no longer required for the register/alignment regulation, since the camera 5 of the image acquisition system 1, which is provided in any case, is used. Moreover, averaging over multiple sheets is possible and special marks can also be used at any arbitrary point on the sheet.
(27) The following is a summary list of reference numerals and the corresponding structure used in the above description of the invention: 1 image acquisition system 2 control computer of the printing press 3 display screen 4 printing press (inkjet or offset) 5 image sensor/camera system 6 user 7 image processing computer 8 grid point for referencing for a color separation 9 calibration points/circular disks of a color separation 10 standard inkjet test printing pattern for a color separation 11 calibration points/circular disks of various color separations in the adapted test pattern 12 adapted inkjet test printing pattern of various color separations 13 alignment register errors in printing bars without corrected tilting 14 alignment register errors in printing bars with corrected tilting 15 printing bar of a printhead 16 well-adjusted alignment 17 moderate alignment deviation 18 large alignment deviation 19 offset color control strips having integrated calibration points/circular disks