EP2774764A1 - Method for generating a printed image made up of sections on a printed material using two ink jet printing heads - Google Patents

Abstract

The method involves selecting a data strip extending in X-direction of width (DY) to a position (Y1) of a print image. The data of the print image in the selected strip is examined based on presence of a data field of expansion dx in X-direction and dy in Y-direction, in which edge detection is carried out in a measurement field (9'). The data field is selected, and X-coordinate and Y-coordinate of the data field are detected. Print images (2, 2a, 2b) are generated. Y-position of two printheads (3a, 3b) in the detected edge is corrected.

Description

The present invention relates to a method having the features of the preamble of claim 1.

The invention is in the technical field of printing with inkjet printheads (so-called "inkjet").

The known state of the art in this technical field includes, for example US 7,955,456 B2 , from which the inkjet printing of blister packs is already known. These have a substantially two-dimensional to be printed sealing film and are conveyed linearly. Despite high production speeds conveying and printing is therefore usually easily possible. Far more difficult is the printing of three-dimensionally shaped bodies with curved outer surfaces in space. In both cases, it may also be necessary to print areas wider than the width of a single printhead. For the printing of packaging, such as bottles in bottling plants, increasingly inkjet printing systems are used. Individual printheads or modules of such systems have, for example, a width of about 50 to about 100 mm, depending on the type and manufacturer. Wider printing systems are therefore constructed by a number of individual modules by stringing together (so-called "stitching") to achieve the required printing width, resulting in inevitable joints.

At these junctions, the individual modules (eg Xaar 1001) must be adjusted to each other (x, y and angle) so that the connection of the modules to each other is not visible in the print as a stripe in the print image: Overlapping or underlapping of the modules manifests itself as unwanted dark (too much ink) or light streaks (too little ink). This adjustment must be carried out both directly after the final assembly and after every service case with the replacement of a module in order to ensure a connection-free print image to ensure. Furthermore, such an adjustment may also be necessary during operation, if the register changes, for example due to thermal influences.

The resolution of currently commercially available modules is between 300 and 600 dpi, which corresponds to 80 to 40 μm. To avoid visibility of a module connector, mispositioning of a printed line in the printed image in the "stitching" area between two modules should not exceed half the line width (40 or 20 μm in the examples above). Such position errors can definitely arise due to temperature influences during operation.

The known prior art further includes the WO 2011/011038 A1 , Which discloses a method of avoiding print defects, wherein during a setup, a test image is printed first in this test image is a so-called "random error" is measured and the print heads with corresponding corrected (called "masked") are driven data, ie the correction takes place at the level of the data. In the process, it is recognized that dark areas in the print image are more susceptible to visible errors than bright areas, which is why image content may be corrected and why an LUT architecture is chosen in which various weighting factors can be set (eg multipliers, the consider different speeds of the substrate). During subsequent production, the print image is monitored and the intensity of the masking is dynamically adjusted as the print density changes (so-called "print density"). A problem with the use of test images may be that - depending on the stability of the system - several test images may have to be printed and measured successively (iteratively), which takes time and produces waste. The document does not name any mechanical adjustment of the heads. Above all, it is not described here that only the printed image or its evaluation could be used for the correction and therefore printing of test images could be avoided.

Against this background, it is an object of the present invention to provide a method which is improved over the prior art and which makes it possible to Reduced quality when printing with multiple inkjet printheads due to so-called stitching errors (errors caused by stacking multiple printheads too far or too close to each other) and avoiding the upstream, or iterative printing of test images and their To refrain from measuring.

An inventive solution to this problem is a method having the features of main claim 1. Advantageous developments of this invention will become apparent from the accompanying dependent claims and from the description and the accompanying drawings.

An inventive method for generating a composed of sections printed image on a substrate with two inkjet printheads, wherein the printed image in the y direction of a first and a second printed image section is substantially composed at a y-coordinate Y1 and wherein a first printhead the first portion and a second printhead produce the second portion, includes the steps of selecting an x-directional data strip of width DY about location Y1 in the data of the print image, examining the data of the print image in the selected stripe presence of a data field of the extension dx in the x-direction and dy in the y-direction, in which an edge detection is possible, selecting the data field and detecting the x-coordinate x1 and the y-coordinate y1 of the data field, Generating the print image, recording an image of a field correlated with the data field on the substrate with a Camera, the image being taken at the position x1, y1 and having at least the size dx * dy, performing edge detection in the data of the image of the measuring field, and - correcting the y-position of at least one of the two printing heads in the case of a detected edge ,

The method according to the invention advantageously makes it possible to avoid quality losses when printing with a plurality of inkjet printheads due to so-called stitching errors and to dispense with the upstream, possibly iterative printing of test images and their measurements. So the process does not need any additional printouts and data for this, but works directly with the printed image or the information or data obtained from it. An elaborate creation of matching test images and the printing of the same and the associated unwanted generation of waste can therefore be omitted. Instead, the method according to the invention finds suitable locations on the printing material for a camera inspection in the real printing data. The method, since it can be carried out predominantly computer-aided, can be carried out very quickly and the correction of the printhead positioning can take place correspondingly fast. In addition, such corrections can be made during printing, so that even small misalignments are always detectable and correctable. Therefore, the method is also ideal for a closed control.

A preferred development of the method according to the invention can be distinguished by the fact that in the case of edge detection a Sobel filter (synonym: Sobel operator) is used. The Sobel filter calculates the first derivative of the gray value curve, while simultaneously smoothing orthogonal to the direction of the derivative. The underlying algorithm uses a convolution using a matrix, which generates a gradient image from the original image, which displays high frequencies in the original image with gray values. The areas of greatest intensity are where the brightness of the original image changes the most and thus have the largest brightness edges. Therefore, a threshold value comparison is usually carried out after the convolution with the Sobel operator.

A preferred development of the method according to the invention can be distinguished by the fact that, when examining the data of the printed image, it is examined whether a data field having an average color density of about 50% to about 70% is present. Lower or higher values (about 30% to about 90%) are useful, although not preferred. Even a mean color density of 100% can still be used: an overlap can not be determined in a measurement field with 100% color density, but an underlap is all the more visible through the resulting bright line. Medium color density values between about 50% to about 70% are therefore preferred because both overlap and underlap can be reliably detected. A preferred development of the method according to the invention can be distinguished by the fact that, when examining the data of the printed image, it is examined whether the data field has a homogeneous, ie uniform (and not heterogeneous or too inhomogeneous) color density. In such a data field, overlapping and underlapping would result in well detectable dark or light lines along with their detectable edges.

A preferred further development of the method according to the invention can be distinguished by the fact that, when examining the data of the printed image, it is examined whether the data field has no or only a slight to maximum mean gradient of the color density or of the corresponding gray value, in particular a profile for the image the slope of the color density progression or the corresponding gray value curve of the printed image in the field (within the data strip) is between about 0 (no gradient) and about 0.5 (range of the maximum average gradient), preferably between about 0.05 and about 0 , 25 (range of average), and more preferably about 0.1.

A preferred embodiment of the method according to the invention can be characterized in that the selection of the data strip, the examination and selection of the data field occur before the beginning of the printing and the coordinates of the data field or the correlated measuring field detected in a data field. Memory for taking the image are kept. In such a process, often referred to as "preflighting", the data required for the inspection (image acquisition and image analysis) can be obtained by computational means. Calculation time does not matter much, because the calculations are essentially completed before printing and inspection begin.

A preferred further development of the method according to the invention can be distinguished by the fact that the method is carried out as a closed-loop control by repeatedly recording an image, performing an edge detection in a control loop and performing a print head correction. In this way it is possible, even small and / or slowly arising or changing misalignments without significant delay to recognize and correct, so that all the print images of a print job with a variety of the same or changing print images of excellent quality.

A preferred development of the method according to the invention can be distinguished by the fact that the method is carried out separately for producing multicolored printed images for at least two color separations. Preferably, in the known four-color printing with the colors cyan, magenta, yellow and black (CMYK), each so-called color separation C, M, Y and K are treated separately according to the invention and thereby subject to any necessary correction, preferably by way of a closed control.

A preferred embodiment of the method according to the invention can be characterized in that the position correction of the print head is effected by a motorized adjustment. For this purpose, a fast-reacting actuator is preferably used, so that necessary position corrections of the print heads can be carried out essentially without any time delay and thus faulty printed images can be substantially avoided.

The abovementioned method according to the invention can also be described by the following synonymous main claim or its combination of features: A method for inkjet printing, wherein a camera takes an image of a measuring field on the printing material, a computer determines a misalignment of a printhead by image processing and controls the lateral position of the printhead and thereby adjusted by motor.

The invention as such and structurally and / or functionally advantageous developments of the invention will be described below with reference to the accompanying drawings with reference to a preferred embodiment.

Figur 1

Eine schematische Darstellung eines bevorzugten Ausführungsbeispiels eines Druck-Systems bei der Durchführung eines bevorzugten Ausführungsbeispiels des erfindungsgemäßen Verfahrens;

Figur 2

Eine schematische Darstellung des Druckbildes bzw. seiner Repräsentation in Form einer x-y-Datenmenge; und

Figur 3

Eine schematische Darstellung von beispielhaften Kamerabildern.

The drawings show:

FIG. 1

A schematic representation of a preferred embodiment of a printing system in carrying out a preferred embodiment of the method according to the invention;

FIG. 2

A schematic representation of the printed image or its representation in the form of an xy data set; and

FIG. 3

A schematic representation of exemplary camera images.

FIG. 1 shows a substrate 1 with a printable area 2, which is composed of two sections 2a and 2b. Section 2a is printed by an ink jet print head 3a and section 2b by an ink jet print head 3b. In the y-direction, the two sections collide at the point Y1. The two printheads overlap in a region DY, which can also be referred to as stitching region. In this area, due to inaccurate positioning (misalignment) of the printheads in the lateral y-direction, unwanted light (too large a distance between the printheads) or dark lines (too much overlap of the printheads) in the printed image may occur. Although a newly installed at the customer printing device is initially adjusted highly accurate, but it can later, also due to thermal influences, come to very fine and only slowly changing misalignment, despite their small deviations from the setpoint but visible and therefore disturbing effects (eg the mentioned lines) in the printed image. Such lines should be avoided according to the invention.

Print head 3b is adjustable in the y-direction. As drive for the adjustment, an actuator 4 is provided. A camera 5, which can be adjusted in the x-direction, is directed onto the printing substrate 1 and can display images 11 (see FIG FIG. 3 ) of areas of the substrate. These images are made available to a computer 6 of a control device.

FIG. 2 shows the print image 7 to be printed in the printable area 2. The print image exists as a file containing an xy-matrix of print image data, eg a so-called bitmap file. In the data matrix, an x-directional strip 8 is selected which is substantially equal in width to the region DY and preferably symmetrical to the y position of the point Y1. This area or the print image data lying in it is selected because it is to be expected that said unwanted light or dark lines. It can also be one wider strip are selected, in which the strip is 8. The selected data in this strip will be examined below.

Examining the selected print image data is done as follows: It is examined in the computer 6 using a corresponding program, whether in the strip 8 at any point (with the coordinates x1 and y1) a field 9 of the size or expansion dx in x direction and dy in the y direction is present, as the measuring field 9 '(see FIG. 1 ) can serve on the substrate 1. The program examines the data in the field 9 as to whether they are suitable for edge detection on the basis of known image processing steps, for example for the application of a digital edge filter such as the so-called Sobel filter. The field 9 is suitable if, for example, it has a substantially homogeneous gray value. "Gray value" in this context can also be understood as a single color value of the color values C, M, Y or K in the conventional production of four-color prints. For CMYK printed images, the examination for presence of a field is therefore carried out separately for each color separation. In fields with a very inhomogeneous gray value or in fields which have characters, for example, because of the high spatial frequencies of the gray value curve, edge detection alone would not be sufficient to detect an existing line. Therefore, such fields are not selected.

1. A) Es weist eine mittlere Farb-Dichte von etwa 50% bis etwa 70% auf.
2. B) Es weist einen im Wesentlichen homogenen Farbdichteverlauf bzw. entsprechenden Grauwertverlauf oder einen nur leichten bis maximal mittleren Farbdichteverlauf bzw. entsprechenden Grauwert-Verlauf auf, der durch das eingesetzte Kanten-Filter aufgrund der zu geringen Ortsfrequenz des Verlaufs nicht erkannt wird. Für den Verlauf kann gelten: die Steigung des Farbdichteverlaufs bzw. des entsprechenden Grauwertverlaufs des Druckbilds im Feld 9 (innerhalb des Datenstreifens 8) liegt zwischen etwa 0 (kein Verlauf) und etwa 0,5 (Bereich des maximal mittleren Verlaufs), bevorzugt zwischen etwa 0,05 und etwa 0,25 (Bereich des mittleren Verlaufs) und besonders bevorzugt bei etwa 0,1.
3. C) Es muss eine solche Lage x1, y1 und Ausdehnung dx * dy haben, dass es komplett im Streifen 8 liegt und es muss genügend groß sein, dass das digitale Filter in seiner vollen Filtergröße über die Stelle Y1 geschoben werden kann. Beispiel 1 (Idealfall): Das Sobel-Filter habe eine Größe von 5 * 5 Pixeln und jedes Pixel eine Größe von 10 * 10 µm. Für DY ergibt sich dann 2* 5 * 10 µm = 100 µm minimale Breite. Beispiel 2 (Praxisbeispiel): Die Lage der Stoßstelle hat bereits eine Ungenauigkeit von etwa 100 µm. Hinzu kommen Ungenauigkeiten hervorgerufen durch die Beabstandung der eingesetzten Kamera vom Ort des Druckens und vom Ort der Trocknens der Druckfarbe oder hervorgerufen durch thermische Effekte an den mechanischen Halterungen der Druckköpfe und der Kamera. Die Breite des Streifens DY wird daher in der Praxis etwa 2 mm oder mehr betragen. Eine Kamera mit 10 mm Bildkreis kann einen solchen 2 mm breiten Streifen problemlos erfassen, der Streifen könnte sich sogar noch mehrere Millimeter seitlich hin- und herbewegen. Das Sobel-Filter habe eine Größe von 5 * 5 Pixeln und jedes Pixel eine Größe von 10 * 10 µm. Der Streifen 8 weise eine Breite von etwa 2 mm auf. Das Sobelfilter lässt sich in diesem Praxisbeispiel somit in seiner vollen Breite in dem Streifen verschieben.

Field 9 is particularly suitable if the following three criteria are fulfilled:

1. A) It has an average color density of about 50% to about 70%.
2. B) It has a substantially homogeneous Farbdichteverlauf or corresponding gray value course or only a slight to maximum average Farbdichteverlauf or corresponding gray value curve, which is not recognized by the edge filter used due to the low spatial frequency of the course. For the course, the slope of the color density progression or of the corresponding gray value profile of the printed image in the field 9 (within the data strip 8) is between about 0 (no gradient) and about 0.5 (range of the maximum average gradient), preferably between about 0.05 and about 0.25 (range of average), and more preferably about 0.1.
3. C) It must have such a position x1, y1 and extension dx * dy that it lies completely in the strip 8 and it must be sufficiently large that the digital filter in its full Filter size can be pushed over the point Y1. Example 1 (Ideal case): The Sobel filter has a size of 5 * 5 pixels and each pixel has a size of 10 * 10 μm. For DY, this results in 2 * 5 * 10 μm = 100 μm minimum width. Example 2 (practical example): The position of the joint already has an inaccuracy of about 100 μm. In addition, inaccuracies caused by the spacing of the camera used from the place of printing and the place of drying of the ink or caused by thermal effects on the mechanical mounts of the printheads and the camera. The width of the strip DY will therefore be about 2 mm or more in practice. A camera with a 10 mm image circle can capture such a 2 mm wide strip easily, the strip could move back and forth even several millimeters laterally. The Sobel filter has a size of 5 * 5 pixels and each pixel has a size of 10 * 10 μm. The strip 8 has a width of about 2 mm. The Sobelfilter can thus be moved in its full width in the strip in this practical example.

If such a field 9 is found in the strip 8, then this is selected for the further progress of the method according to the invention and thus to a measuring field 9 '. If several useful fields, e.g. Fields 9 and 10 are found, then among these the most useful may be selected, e.g. the one that best meets the above criteria A to C. The coordinates (x-coordinate x 1 and y-coordinate y1) of this selected field are detected and are thus available for further steps.

The selection of the data strip, the examination and selection of the data field can preferably take place before the beginning of the printing and the coordinates x1 and y1 of the data field 9 or the correlated measuring field 9 'detected in this case can be stored in a data memory 6 'are provided for the recording of the image 11 with the camera 5.

The camera 5 now takes a picture 11 (see FIG. 3 ) of the measuring field 9 '(see FIG. 1 ) on. For this purpose, the camera, for example, by an adjustment, be brought into a position that allows the recording of the image. However, if the camera is already positioned so that the strip of width DY passes underneath, then the camera only needs it Moment, in which the measuring field in the recording area of the camera passes. For this purpose, the coordinates x1 and y1 of the selected measuring field can be used, which allow for knowledge of the substrate dimensions or the position of the print image 2 on the substrate and the conveying speed of the printing material 1, a calculation of the triggering time of the camera. The recorded image should have at least the size dx * dy, so that an edge detection can be performed in the image.

The data of the captured image 11, e.g. as a bitmap, the computer 5 are provided. The computer or a program running on it now carries out an edge detection, wherein e.g. the above-mentioned Sobel filter is used.

If there is a substantial misalignment of at least one printhead, the edge detection will yield an edge in the data of the image 11 of the measurement field 9 'as a result. In this case, a threshold value comparison can also be carried out: If a given threshold is exceeded, an undesired overlap of the print heads would be detected. If it falls below a given, different threshold, an undesirable underlap, i. too large a distance of the printheads, detected. Depending on the present case (overlap or overlap), the direction of the motorized adjustment of at least one of the print heads is determined.

FIG. 3 shows by way of example three such recorded images 11 of the selected measuring field 9 '. In the left image it can be seen that the two print heads 3a and 3b are exactly aligned with one another and therefore do not produce a visible line at the joint Y1 of the print heads. The picture therefore shows a homogeneous gray value. In the middle picture it can be seen that the two print heads overlap too far and therefore create a dark line at the joint Y1. The picture on the right shows the case where the two printheads are too far apart and therefore produce a bright line in the print image at joint Y1.

If the performed edge detection one of the two in the left or right image of the FIG. 3 detected cases, the computer 6 will send a signal to the actuator 4, which performs a corrective adjustment of the print head 3b. Alternatively, the print head 3a or both printheads can be adjusted. The adjustment takes place in any case in the y-direction and the value of the adjustment correlates with the width of the detected line.

Since the adjustment of the print head 3b leads to a change in the print image 2 and the width of the in FIG. 3 In this case, a renewed recording of an image 11 of the measuring field 9 'in one of the subsequent prints and its renewed evaluation by edge detection can be used to form a closed control loop. In the context of this regulation or the described sampling control system, a measuring field 9 'is repeatedly or even continuously selected and an image taken of it, an edge detection is performed in the data of the image and causes a compensating adjustment of the printhead in the presence of an edge. The measuring field can also be selected only once, preferably before the beginning of printing, and then used again and again during monitoring for the monitoring. For the same prints a fixed measuring field is an advantage, but with changing printing it can be an advantage to redefine the measuring field for each changed print. As a result, the distance of the last nozzle (or nozzle row) of a first print head to the first nozzle (or nozzle row) of a second, adjacent print head is ultimately used as the controlled variable. This distance can be determined from the width of the line (light or dark) found during edge detection.

If CMYK images are printed, it is advantageous to carry out the described process separately for each color separation, that is to say for C, M, Y and K, ie to determine a data field 9 for each color separation and a measurement field 9 'therefrom. The measuring fields for the different color separations do not have to have the same xy coordinates, ie the camera 5 can record images 11 located at different locations from the printed image on the printing substrate 1. The edge detection can then also be carried out separately for each color separation and the results, ie the control values for the motorized adjustment, are assigned to the individual print heads for the different colors CMYK fed. It may therefore be that only one color separation, several or even all color separations or the associated printheads undergo corrections. In order to take pictures of the different color separations, the camera can be equipped with corresponding, preferably automatically changeable color filters or it can be provided to use corresponding, preferably automatically changing illumination means. If no data fields 9 in a strip 8 which can be used for edge detection can be determined for one of the color separations, then the two print heads 3a and 3b concerned can be excluded from the regulation, because then misalignments of the relevant print heads are not to be expected lead to a visible disturbance due to line formation. If the same applies to all joints of printheads in a color separation, then the entire color separation can be taken out of the scheme.

LIST OF REFERENCE NUMBERS

1

substrate

2

Printable area

2a, 2b

sections

3a, 3b

Inkjet printheads

4

actuator

5

camera

6

calculator

6 '

Data storage

7

Printed image (data)

8th

Strip (data)

9

Field (data)

9 '

measuring field

10

Field (data)

11

image

Y1

Joint of the printheads

DY

Width of the strip

x1

x-coordinate of the measuring field

y1

y-coordinate of the measuring field

dx

Width of the measuring field in x-direction

dy

Width of the measuring field in the y-direction

Claims (10)

A method for producing a composite of sections printed image on a substrate with two inkjet printheads, wherein the printed image (7) in the y-direction of a first and a second printed image section (2a, 2b) substantially at a y-coordinate Y1 composed and wherein a first print head (3a) generates the first portion (2a) and a second print head (3b) generates the second portion (2b), comprising the steps of: Selecting an x-directional data strip (8) of width DY around the location Y1 in the data of the printed image (7), - Examining the data of the printed image (7) in the selected strip (8) for the presence of a data field (9) of the extension dx in the x-direction and dy in the y-direction, in which an edge detection is possible, Selecting the data field (9) and detecting the x-coordinate x1 and the y-coordinate y1 of the data field (9), Generating the printed image (2, 2a, 2b), - Taking an image (11) of a correlated with the data field (9) measuring field (9 ') on the substrate (1) with a camera (5), wherein the image (11) at the location x1, y1 is recorded and at least the size dx * dy, Performing an edge detection in the data of the image (11) of the measuring field (9 '), and - Correction of the y-position of at least one of the two print heads (3a, 3b) in the case of a detected edge. Method according to claim 1,
characterized,
that a soot filter is used in edge detection. Method according to claim 1 or 2,
characterized,
that a threshold comparison is carried out at least the edge detection. Method according to claim 1,
characterized,
that is examined at the examining the data of the print image (7), whether a data field (9) is present having an average color density of about 50% to about 70%. Method according to claim 1,
characterized,
that, when examining the data of the print image (7), it is examined whether the data field (9) has a homogeneous color density Method according to claim 1,
characterized,
in that when examining the data of the printed image (7), it is examined whether the data field (9) has no or only a slight to maximum average color density or the corresponding gray value, in particular a curve for the following: the slope of Farbdichtverlaufs or the corresponding gray value curve of the printed image (7) in the field (9) is between about 0 and about 0.5, preferably between about 0.05 and about 0.25 and more preferably at about 0.1. Method according to claim 1,
characterized,
in that the selection of the data strip (8), the examination and selection of the data field (9) take place before the beginning of the printing and the coordinates of the data field (9) and the correlated measuring field (9 ') acquired in one Data memory (6 ') for receiving the image (11) are kept. Method according to claim 1,
characterized,
in that the method is carried out as a control by repeatedly recording an image (11), performing edge detection and performing printhead correction in a control loop. Method according to claim 1,
characterized,
that the method is carried out separately when producing multicolored printed images for at least two color separations. Method according to claim 1,
characterized,
that the position correction of the print head (3a, 3b) is effected by a motorized adjustment.

Images