AU2009298228A1

AU2009298228A1 - Method for generating a detection signal, and detecting device

Info

Publication number: AU2009298228A1
Application number: AU2009298228A
Authority: AU
Inventors: Unal Dogan; Matthias Kuhl; Holger Schnabel; Stephan Schultze
Original assignee: Robert Bosch GmbH
Current assignee: Robert Bosch GmbH
Priority date: 2008-10-02
Filing date: 2009-09-01
Publication date: 2010-04-08
Also published as: WO2010037454A1; JP2012504505A; DE102008049908A1

Description

Translation from German WO 2010/037454 Al PCT/EP2009/006312 Method for Generating a Detection Signal, and Detecting Device Description The present invention relates to: a method for generating a detection 5 signal; and a detection device. Although the following refers mainly to printing presses, the invention is not restricted to that field, but rather is intended for all types of processing machines in which a web of material is processed. The invention can, however, be used in particular with printing machines such as e.g. 10 newspaper presses, jobbing presses, gravure printing machines, package printing machines, and securities printing machines; and with processing machines such as e.g. bag-making machines, envelope-making machines, Translation from German WO 2010/037454 Al PCT/EP2009/006312 and packaging machines. The web concerned may be made of paper, fabric, cardboard, plastic, metal, or rubber, and may be in sheet form, etc. Prior Art In multicolour printing, e.g. in rotary printing machines, the individual 5 colour separations (particularly for cyan, magenta, yellow, and black) are applied, in successive printing units, to a substrate in the form of roll stock that is conducted through the printing-unit continuously. It is crucial for print-quality that the images for each colour be printed exactly on top of one another. This superimposing of the printed images is termed 10 registration. In addition to the actual printed image, each printing unit also prints registration measurement marks, in the form of e.g. crosses, triangles, etc., for adjusting the individual printing units relative to one another. On the basis of these marks, any misalignment of the printed images relative to one another can be detected online by an optical is measuring system. In rotary printing systems, this measurement system is generally part of a control system, known as the registration control system. The registration control system intervenes in the printing process by means of suitable actuators, adjusting for any deviations in registration Translation from German WO 2010/037454 Al PCT/EP2009/006312 detected by the optical measuring system (detecting-device; registration sensor). In particular, the actuators can alter the length of the substrate between successive printing units, so that the images from successive printing-units are superimposed on one another. 5 Contrast sensors, colour sensors, and cameras are used as detection devices. They typically operate in reflection. The web of material is illuminated constantly or pulse-wise by a suitable light source (white or coloured light), and the reflected light is detected by the sensor and evaluated. 10 Before a scan with contrast sensors or colour sensors is performed, a "teach" process normally occurs, in which a signal threshold is learned whereby it is possible to distinguish between "mark" and "no mark". The point in time, during scanning, at which the signal being evaluated passes above or below the signal threshold is taken as the moment of contrast 15 transition. Contrast sensors known in the art are designed as purely total-intensity sensors, and so are "colour-blind". Usually, the web is irradiated with white Translation from German WO 2010/037454 Al PCT/EP2009/006312 light, and the reflected intensity is measured with a photodiode. Such a sensor is shown in DE 10 2004 021 597 Al. Modern sensors and detection devices are capable of illuminating the web with different colours (e.g. red, green, blue, etc.) in a teach process, to obtain optimal contrast for the 5 particular colour of the printed marks. However, there is a problem with this, in that it is often not possible to find an illuminating-colour that will give optimal contrast for all the marks on the material. In that case, it will be necessary to use a number of sensors, thereby increasing the cost, or to use colour sensors, but these are more expensive than pure contrast 10 sensors and have other drawbacks. Known colour sensors are typically designed to evaluate printed marks of different colours. These sensors usually contain a source of white light for illuminating the web of material. The reflected light is captured and broken down into spectral components, typically by means of suitable colour is filters. Alternatively, sequential illumination and detection using light of different colours (usually red, green, blue) is also known in the art. There are three output signals, each corresponding to the intensity of one colour channel. In the teach process, the output signal chosen for mark detection Translation from German WO 2010/037454 Al PCT/EP2009/006312 is the one that gives the greatest difference in intensity between "mark" and "no mark". However, this has the disadvantage that, with certain combinations of registration mark colour and background colour, only a poor signal-to-noise ratio is achievable. 5 It is therefore desirable to create a method for generating a detection signal, and a detection device, that will make it possible to reliably detect registration measurement marks. Given this background, the present invention proposes a method for generating a detection signal, and a detection device, with the features of 10 the independent claims. Advantageous further developments are the subject of the dependent claims and the description below. Advantages of the Invention The inventive solution makes it possible to provide a detection device or colour sensor that, with a few computational operations, can provide a is high-resolution colour detection signal. The colour sensor can, in particular, be implemented as a True Colour XYZ Sensor. In this way, it is possible, in particular, to generate a high-resolution colour and intensity Translation from German WO 2010/037454 Al PCT/EP2009/006312 signal as an xyY signal. In this way it is possible, in particular, to indicate a continuous colour error with respect to the background. Search algorithms can be applied to the detection signal, in order to first locate the registration measurement marks roughly and then interpolate 5 them precisely. It is possible to use filter characteristics that substantially correspond to the human eye: this can be achieved by using the XYZ colour space. By using at least two, or preferably three, colour signals to produce the detection signal, the detection signal can be adjusted optimally to the registration measurement mark that is to be detected. 1o The invention is particularly useful for detecting and recognising registration measurement marks in printing machines. To generate the detection-signal based on the two or more first colour signals, it is advantageous to initially generate at least two second colour signals by transforming the two or more first colour-signals, and then 15 generate the detection signal based on the two or more second colour signals. Preferably, the first and/or second colour-signals and/or the Translation from German WO 2010/037454 Al PCT/EP2009/006312 detection-signal can be outputted - for which, analogue or digital outputs such as e.g. fieldbus interfaces can be used. It is expedient to generate the detection signal as an analogue voltage waveform, particularly a curve of discrete-time measurement values in 5 which time-spaced detection values are produced and outputted at e.g. an interface or output. It will be appropriate, in this regard, to detect a registration measurement mark located on the material being processed - particularly printing-stock - by comparing the detection signal with a pre-settable threshold value; after which, a switching signal in particular 10 can be outputted that has a flank corresponding to the moment when the detection signal becomes greater or less (depending on the setting) than the pre-set threshold value. It is beneficial for the detection device to be able to be preset with a number of threshold values, so that the detection device can detect a 15 number of registration measurement marks located on the material being processed. It is possible, in particular, to select a particular, suitable, threshold value for each different registration mark (different colour, or contrast). It is also advantageous if the two or more first or second colour Translation from German WO 2010/037454 Al PCT/EP2009/006312 signals go into the detection signal with a weighting that is presettable. It is also advantageous if the detection device can be preset with a number of weightings for detecting a plurality of registration measurement marks on the material being processed. 5 The one or more threshold values and/or the one or more weightings may be user-presettable or user-alterable, in particular externally by means of parameters, or automatically by a higher-level control system. Also, automatic presetting can be performed within the detection device itself, by suitable means such as a computing unit. Automatic or manual presetting 10 can be performed repeatedly, to keep the detection signal as optimal as possible throughout the processing of the web of material. When automatically or manually presetting the one or more threshold values and/or weightings, it is preferable to take into account or maximise a signal-to-noise ratio; or to maximise the amplitude of the detection signal. 15 The weighting factors for the decision rules are preferably adjustable for each mark separately. If, for example, a black mark and a colour mark with low contrast on a white background are to be detectable, then the weighting for the black mark should be set greater for intensity (Y) and the Translation from German WO 2010/037454 Al PCT/EP2009/006312 weighting for the low-contrast colour mark should be set greater for colour recognition (xy). It is also appropriate to determine the weighting factors by means of signal-to-noise ratios (in particular, automatically - i.e. autonomously by 5 the sensor, or automatically by a controller or a registration control system). Depending on the noise component of the mark concerned, it may be advantageous to optimise the weighting factors so as to obtain the maximum possible signal-to-noise ratio. This will facilitate the subsequent determination of the mark flank, and will make that determination more 10 robust. In addition, the factors can be optimised on the basis of the current level of the combined signal for the mark concerned. The aim of this optimisation is to obtain the greatest possible difference in level between the background and the mark. 15 It is appropriate to interpolate the detection signal, the two or more first colour signals, and/or the two or more second colour signals, when they are available as a sequence of discrete-time detection-values. In this way, Translation from German WO 2010/037454 Al PCT/EP2009/006312 an accurate comparison with the pre-set threshold value can be made, whereby the registration mark can be more accurately determined. This temporal interpolation can, for example, be linear, cubic, sinusoidal, polynomial, Gaussian, etc. Instead of comparing with a threshold value, 5 the registration mark can also be determined by means of a detected inflexion point. In the prior art, comparing the measurement signal with the threshold value is only ever performed at the time of measurement. The resolution of the registration measurement marks therefore depends directly on the sampling frequency. With the preferred interpolation 10 method, this dependence can be eliminated, making it possible to sample more accurately and to achieve increased temporal resolution. On the basis of the evaluated signal, the sensor determines a point in time at which the mark('s flank) occurs. This can be converted into a registration difference in the sensor or in the controller algorithm. The accuracy and 15 resolution of temporal detection affects the quality of the control process, and should preferably be maximised. By interpolation- and filtration mechanisms, a possibly noisy total signal can, accordingly, be improved for analysis. Of the total signal, it is basically only the signal flanks (flanks of the mark) that are of interest, and therefore the total signal can be Translation from German WO 2010/037454 Al PCT/EP2009/006312 examined more closely in the region of the flank, in an evaluation step. For example, using best-fit algorithms or the least error squares method, a function can be fitted to the measurement values. If mark-quality should now change in the course of the process, the detection signal will also 5 change (e.g. decrease in amplitude) due to the high resolution of the XYZ sensor. The profile described by the curve of the detection signal over time will, however, remain substantially the same. In a particularly preferred embodiment of the invention, the detected light is decomposed into three spectral bands, e.g. RGB or XYZ. In this way, 10 particularly-advantageous full colour information can be provided. In the prior art, mainly so-called "RGB sensors" are used. In this regard, one can also talk of the red, blue, and green channels. The (colour) filter functions implemented are manufacturer-specific, and divide the visible light spectrum into three bands. In this connection, the term (device-specific) is "RGB" should not be confused with one of the defined RGB colour spaces, sRGB for example. Rather, it relates to a manufacturer/sensor-specific spectral breakdown.

Translation from German WO 2010/037454 Al PCT/EP2009/006312 It is beneficial if the three second colour signals are XYZ or xyY signals. Since, as just mentioned, RGB sensors provide manufacturer-specific RGB values, a standardised colour-value output will be advantageous for applications requiring defined colour determination (such as certain s printing operations). This can be achieved in particular by converting the first (sensor-specific RGB) colour signals into second (normalised CIE) colour signals, in particular the CIE normalised theoretical primary colours X (red), Y (green), and Z (blue). The XYZ colour space is device-independent. CIE1931 describes filter 10 characteristics that are intended to correspond, as far as possible, to the human eye. Numerous other colour spaces can be traced back directly to the XYZ colour space or can be derived from it mathematically. One of these colour spaces is the xyY colour space (two-dimensional CIE chromaticity diagram), which can be calculated from XYZ. Another colour 15 space is the CIELab colour space. This is more complex to convert, but has the great advantage of providing a single number AE describing a colour error that also corresponds to human visual perception. AE can thus be used in particular as a detection signal. Conversion to XYZ, xyY, etc., is Translation from German WO 2010/037454 Al PCT/EP2009/006312 performed by matrix or tensor multiplication, in a manner known in the art. The CIE chromaticity diagram contains all possible colours in a tongue shaped figure on an xy coordinate system. Intensity Y (brightness darkness) is the third component used for defining the colour. 5 With prior-art RGB colour sensors, a sensor-specific correction matrix has to be applied to the sensor values in a higher-level control system in order to obtain XYZ signals. With the form of embodiment of the present invention now being described, however, these signals are advantageously - outputted directly, thereby saving on computing time in 1o the control system. The three information-items x, y, and Y can now be combined to form the detection signal, particularly using the above-described weighting: a single combined number for a mark can be formed from the two values x and y. This number could, for example, be the root of the sum of the squares of x 15 and y. This combined colour information can be used for evaluating the printed mark (colour evaluation). In addition, the value of Y (intensity) will preferably also be used for evaluating the printed mark. Actually, if only x and y are used, then only a colour evaluation will be obtained; and if only Translation from German WO 2010/037454 Al PCT/EP2009/006312 Y is used, then only a contrast evaluation will be obtained. The disadvantage with just a colour evaluation is, for example, that it cannot distinguish between black, grey, and white, which, in the xy colour space, are on the same point and only have different intensities Y. The 5 disadvantage with just a contrast evaluation is that, with the sensors available today, it will be relatively noisy; and, with very low-contrast marks (e.g. a yellow mark on a gold background), it will be hardly detectable. By combining these two elements - "colour information" and "contrast information" - these disadvantages can be avoided. Combining these two to items of information into a single total information item is affected by the weighting employed. Weighting can be done for example by providing each of the two values with a factor such that the sum of the two factors is 1. This will give a readily-variable control range between "contrast evaluation" and "colour evaluation", and any desired values inbetween. 15 Preferably, a plausibility check of the colour of the detected registration measurement mark will be performed by means of the three first and second colour signals. In particular, it is possible to check whether the colour of the mark is correct in each case, or whether a change has Translation from German WO 2010/037454 Al PCT/EP2009/006312 occurred in the colour - in which event, an error message or warning can be outputted. It is a beneficial to supply to the detecting device an item of information as to the web's position, and use this information and the detection signal to 5 determine a length value pertaining to the registration measurement mark on the web of material. The length value can describe a mark-location or a registration error. In this way, an intelligent detection device can be provided which will relieve a higher-level control system of the burden of determining a registration position or a registration deviation, by taking on 1o that task itself. The length value so determined can e.g. be outputted through a fieldbus interface. Similarly, the information as to the position of the web can be supplied via a fieldbus interface; and in this case, a realtime-capable fieldbus, preferably an Ethernet-based one such as SERCOS Ill, should be used. 1s A detection device according to the present invention has means for implementing a method according to the present invention. Outputs can, in particular, be analogue or digital - e.g. a fieldbus interface. The invention prefers the use of detection devices in the form of colour sensors, which Translation from German WO 2010/037454 Al PCT/EP2009/006312 are employed to advantage when a relatively large number of printing units with many different colours are used. The content of the invention can, however, also extend in principle to the use of colour cameras, without thereby going beyond the scope of the invention. The detection device will, 5 advantageously, comprise an interface for a field bus - preferably a realtime-capable field bus, and particularly an Ethernet-based one such as SERCOS Ill. The detection device can then be integrated into a machine control system, particularly in a printing press, by means of a field bus. Further advantages and embodiments of the invention will emerge from to the description and the accompanying drawings. It will be appreciated that the features mentioned above and those yet to be explained below can be used not only in the particular combinations indicated but also in other combinations, or in isolation, without thereby departing from the ambit of the present invention. 15 The invention is represented schematically in the drawings, which show examples of its embodiment; and it will be discussed in detail below, with reference to the drawings.

Translation from German WO 2010/037454 Al PCT/EP2009/006312 Description of the Figures Figure 1 is a schematic representation of a preferred form of embodiment of a detection device according to the present invention; and 5 Figure 2 is a flow diagram of a preferred form of implementation of a method according to the present invention. Figure 1 is a schematic representation of a preferred embodiment of the inventive detection device 200. The detection device 200 here is designed as a registration mark sensor and is used for detecting registration marks 10 or registration measurement marks on a web 102 of material. The web 101 travels in direction R, relative to the registration mark sensor 200. The registration mark sensor 200 shown here has a source of light, in the form of a white-light LED 201, to illuminate the material being processed. Of course, the material being processed can also be illuminated by a light 15 source that is not part of the detection device 200. The light is reflected from the web 101 and enters the detecting device 200 as a light beam 202.

Translation from German WO 2010/037454 Al PCT/EP2009/006312 The detection device 200 also has a beam splitter or refracting element in the form of a half-silvered mirror 203 in the present instance. This deflects the light beam 202 onto three optical sensor devices, which are photo diodes 204 in the present example. In the drawing, these 5 photodiodes 204 are each provided with a colour filter 205, to divide the reflected light into three spectral components: red, green, and blue. The photodiodes 204 each serve as means of generating a first colour signal R, G, B for the spectral bands. The three first colour signals R, G, B are fed to a means of generating a detection signal S, said means being a 1o computing unit in the present example. The computing unit 206 is designed to serve, at the same time, as a means of generating three second colour signals x, y, Y, which, in the example shown here, are outputted from the detection device 200 through an output in the form of a fieldbus interface 207. Preferably, it is also possible to feed parameters a, 1s b, and c into the detection device 200 through the fieldbus interface, in order to adjust the detection signal S. Taking the supplied parameters into account, the computing unit 206 calculates the detection signal S, which can be outputted, through an analogue output 208, as a sequence of Translation from German WO 2010/037454 Al PCT/EP2009/006312 discrete-time measurement values. The detection signal S can also be outputted through the fieldbus interface 207. The detecting device can also be designed so that the photodiodes 204 are arranged alongside the colour filters 205 or integrated in a chip, and 5 then illuminated by the reflected light without beam-splitters. Or course, a colour-sensitive array of photodiodes or a colour-sensitive CCD sensor can also be used for this. The detection device 200 shown in the drawing is designed for receiving positional data P regarding the web 101, via its fieldbus interface 207 or 10 other positional interface, e.g. a pulse encoder. Based on the position data P received, which may include e.g. the required and/or actual angular position(s) of pressure cylinders, and/or machine angles (so-called leading-axis positions), and thus the required position of the registration measurement marks 102 on the product web 101, the detection device 15 200 can, on detecting the registration measurement marks 102, determine a required position for the registration measurement marks and/or any deviation L from the required position, and supply this result over the Translation from German WO 2010/037454 Al PCT/EP2009/006312 fieldbus interface 207, thereby relieving a higher-level control system of the task of calculating the deviation. Figure 2 is a flowchart of a preferred form of implementation of the inventive process. The process begins in step 300. In step 301, a material 5 being processed is illuminated. In particular, this illumination can be continuous or pulsed. In step 302, the light reflected from the material being processed is detected by an optical sensor device. Of course, it is also possible to work in transmission mode, i.e. with the optical sensor device detecting transmitted light. to In the following step 303, the detected light is split into three spectral bands, namely red, blue, and green. Then, in step 304, the intensities of the three spectral bands are determined, these being the first colour signals. In the preferred form of implementation of the invention, shown in the diagram, xyY signals are then generated in step 305, these being the 15 second colour signals. In step 306, the second colour signals are combined to produce a detection signal. This can, in particular, be done using three externally presettable parameters a, b, c and generating the detection signal as Translation from German WO 2010/037454 Al PCT/EP2009/006312 (ax 2 + by2)% + cY for example. The parameters a, b, c can also be calculated automatically in the detection device, or in a higher-level control system, so as to e.g. provide an optimum signal-to-noise ratio for the detection signal. The 5 detection signal is used to detect the registration measuring marks on the material being processed.

Claims

1. A method for generating a detection signal, by means of a detection device (200) for detecting a registration measurement mark (102) on a material being processed (101), with the following steps: 5 - illuminating the material being processed (101), by means of a light source (201), - detecting the reflected or transmitted light (202) by means of an optical sensor device (203, 204), decomposing the detected light (202) into at least two spectral bands, and determining the intensity of each spectral 10 band so as to generate at least two first colour signals (R , G, B), and - generating a detection signal (S) based on said two or more first colour signals (R, G, B), each of said two or more first colour signals (R, G, B) going into the detection signal (S).

2. A method as claimed in claim 1, wherein - to generate the detection 15 signal (S) based on the two or more first colour signals (R, G, B) - at least two second colour signals (x, y, Y) are generated by transforming the two or more first colour signals (R, G , B), and then the detection signal (S) 23 WO 2010/037454 Al PCT/EP2009/006312 is generated on the basis of said two or more second colour signals (x, y, Y).

3. A method as claimed in one of the above claims, wherein the detection signal (S) is in the form of an analogue voltage wave, which can 5 in particular be outputted from the detection device (200).

4. A method as claimed in claim 3, wherein a registration measurement mark (S) on the material being processed (101) can be detected by comparing the detection signal (S) with a presettable threshold value.

5. A method as claimed in claim 4, wherein the detection device (200) is 10 able to be preset with a number of threshold values for detecting a number of registration measurement marks (102) on the material being processed (101).

6. A method as claimed in any of the above claims, wherein the two or more first or two or more second colour signals (R, G, B; x, y, Y) go into 1s the detection signal (S) with a presettable weighting. 24 WO 2010/037454 Al PCT/EP2009/006312

7. A method as claimed in claim 6, wherein the detection device (200) can be preset with a number of weightings for detecting a number of registration measurement marks (102) on the material being processed (101). 5

8. A method as claimed in any of claims 4 to 7, wherein the setting of the one or more threshold values and/or weightings is performed automatically.

9. A method as claimed in claim 7 or 8, wherein the setting is performed taking in account a signal-to-noise ratio, particularly with a view to 10 maximising the signal-to-noise ratio.

10. A method as claimed in claim 7, 8, or 9, wherein the setting is performed on the basis of maximising the amplitude of the detection signal (S).

11. A method as claimed in any of the above claims, wherein the 15 detection signal (S) for detecting a registration measurement mark on the material being processed (101) is interpolated. 25 WO 2010/037454 Al PCT/EP2009/006312

12. A method as claimed in any of the above claims, wherein the two first or more colour signals (R, G, B) and/or the two or more second colour signals (x, y, Y) and/or the detection signal (S) are interpolated to increase the time resolution of the signal sampling. 5

13. A method as claimed in any of the above claims, wherein the detected light is broken down into three spectral bands.

14. A method as claimed in claims 13 and 2, wherein the three second colour signals are XYZ signals or xyY signals.

15. A method as claimed in claim 13 or 14, wherein, by means of the 1o three first or second colour signals, a plausibility check is performed as to the colour of the detected registration measurement mark (102).

16. A method as claimed in any of the above claims, wherein the detection device (200) is fed an item of information (P) as to the position of the web, and on the basis of this information (P) and the detection signal 1s (S), a length value (L) of the registration measurement mark (102) on the web is determined. 26 WO 2010/037454 Al PCT/EP2009/006312

17. A detection device (200) for detecting a registration measurement mark (102) on a material being processed (101), with: - an optical sensor device (203, 204, 205) for detecting light (202), - means (203, 204, 205) for decomposing the detected light into at least 5 two spectral bands, - means (205, 206) for generating a first colour signal (R, G, B) for each of the two or more spectral bands, - means (205, 206) for generating a detection signal (S) based on the two or more first colour signals (R, G, B), each of said two or more first 10 colour signals (R, G, B) going into the detection signal (S).

18. A detection device (200) as claimed in claim 17, wherein the means (206) for generating the detection signal (S) comprise means (206) for generating at least two second colour signals (x, y, Y) by transforming the two or more first colour signals (R, G, B). 15

19. A detection device (200) as claimed in claim 17 or 18, with means for generating a switching signal by comparing the detection signal (S) with a presettable threshold value. 27 WO 2010/037454 Al PCT/EP2009/006312

20. A detection device (200) as claimed in claim 17, 18, or 19, with at least one output (207, 208) for outputting the following signals: the detection signal (S), the switching signal, the two or more first colour signals (R, G, B), and/or the two or more second colour signals (x, y, Y). 5

21. A detection device (200) as claimed in any of claims 17 to 20, wherein the means for decomposing the detected light into at least two spectral bands are in the form of means (203, 204, 205) for decomposing the detected light (202) into three spectral bands.

22. A detection device (200) as claimed in claims 18 and 21, wherein the 10 means for generating at least two second colour signals are in the form of means (206) for generating XYZ signals or xyY signals.

23. A detection device (200) as claimed in any of claims 17 to 22, with an interface for a fieldbus, particularly a realtime-capable Ethernet-based fieldbus, especially SERCOS Ill.