Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
METHOD AND SYSTEM FOR MONITORING PRINTED MATERIAL PRODUCED BY A PRINTING PRESS
Document Type and Number:
WIPO Patent Application WO/2005/087494
Kind Code:
A1
Abstract:
The invention relates to a method for monitoring printed material, which is produced by a printing press and comprises images printed on a paper web. The method comprises the steps of: determining in a digital original reference values for chosen parameters of the printed material, including in any case the colors, detecting these parameters in a digitized record of the corresponding printed images, comparing the detected values to the reference values, and performing a correction when a difference in these values is found during the comparison. With a view to the comparison, use is made for describing the colors of a system adapted to the human eye, such as CIELab. The chosen parameters can also comprise, in addition to the colors, the location of the images in the printed material and/or the color register of the printed material. The invention also relates to a system for performing this method, which is provided with a device for determining reference values, a device for detecting the parameters, a comparing and correcting device connected to the determining device and the detecting device.

Inventors:
VAN HOLTEN ERIK (NL)
JANSEN MENNO (NL)
Application Number:
PCT/NL2005/000187
Publication Date:
September 22, 2005
Filing Date:
March 14, 2005
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
Q I PRESS CONTROLS HOLDING B V (NL)
VAN HOLTEN ERIK (NL)
JANSEN MENNO (NL)
International Classes:
B41F33/00; (IPC1-7): B41F33/00; B41F33/12
Foreign References:
US6024018A2000-02-15
US5774635A1998-06-30
DE19940879A12001-03-08
DE19855177A12000-06-08
EP0850763A11998-07-01
US5774635A1998-06-30
US6024018A2000-02-15
Attorney, Agent or Firm:
Bartelds, Erik (Sweelinckplein 1, The Hague, NL)
Download PDF:
Description:
METHOD AND SYSTEM FOR MONITORING PRINTED MATERIAL PRODUCED BY A PRINTING PRESS

The invention relates to a method for monitoring printed material as according to the preamble of claim 1. Such a method is known from US 5,744,635. In a printing process, for instance a rotation offset printing process, a large number of identical copies is made on the basis of an original, for instance an illustration or a text. It is of great importance here that the images in the printed material are transferred to the substrate for printing, generally a paper web, as far as possible in colour-true, correctly positioned and consistent manner. "Colour-true" is here understood to mean that the printed colours correspond precisely with the colours of the original which is being duplicated. This is especially important for advertisers, who increasingly use a colour as a mark and therefore wish to see it printed correctly. The correct position is important because it determines the general picture of the printed material, while consistency is important in order to indeed be able to ensure identical copies of the printed material. A problem which occurs here is that the final colour of a printed image is influenced by a large number of variables which cannot be fully monitored during printing. Differences will thereby occur between the original and the printed images. In order to be able to understand the causes of these differences, some insight into the nature of printed materials is required. Printed materials are built up of grid points or dots. The printed image is determined by two quantities, the number of grid points per unit of length, usually expressed in "dpi" (dots per inch), which specifies how fine or crude the image is, and the relation between light and dark, the tonal range. As long as the tonal range takes on values of less than 50%, the image is formed by dark, thus printed dots on a light background, while for higher values use is made of light dots, thus dots not printed in the surrounding, printed background. During the different steps which must be performed during the printing process to form a printed image from an original, there do however occur differences in the size of the grid points; the points grow, designated as "dot gain" or shrink, designated as "dot loss". The final printed image hereby differs from the original. The degree to which this occurs depends on a large number of variables associated with the printing process, and on the original size of the grid points. The finer the printed material is, i.e. the closer together the grid points, the greater the differences are. Furthermore, the differences are often relatively greatest in parts of the printed material where the tonal range amounts to about 50 percent. In practice therefore, printing presses are calibrated before use. Proofs are made with different finenesses and different tonal ranges, and the differences are determined from these proofs. These measured differences are stored in the form of calibration graphs, so-called dot gain curves. These graphs are used to determine the theoretical optimal settings of each printing press when an original with a determined fineness and tonal range is presented. As stated, there are many variables whereby the quality of the final printed material is influenced. Envisaged here are the type of paper used, the water-ink balance, the temperature, possible contamination of the ink or the ink rollers, the pressure between the different ink rollers, vibration occurring in the printing presses and the like. A change in just one of these quantities can result in an increase or decrease in size of the grid points, and therefore in differences in the printed colours. Because it is almost impossible to hold all these quantities constant during the printing process, different methods have been developed to control these quantities on the basis of inspection of the colours such that the quality remains constant. There are thus so-called open loop control methods, wherein random samples of the printed material are monitored using hand-held measuring devices. These hand measuring devices monitor a separately co-printed colour bar. When differences are detected an operator can change the settings of the printing press. A drawback of this control method is the discontinuous and even informal nature thereof. Between two consecutive random samples a large quantity of printed material of substandard quality can be produced. This method moreover requires the presence of an often expensive operator. Fully closed control methods have therefore been developed, wherein measuring and control systems monitor the quality of the printed material fully autonomously and adjust the settings of the printing press when differences are detected. These known systems are also based on a co-printed colour bar. There are two major trends here. Known in the first instance are systems wherein the colour bar consists of areas which are completely covered with the basic colours C (cyan), M (magenta), Y (yellow) and K (black), whereby the colour density (thickness of the ink layer) can be measured. In improved systems the colour bar also comprises grey areas, including a black area with 50% tonal range and an area with cyan, magenta and yellow in quantities such that black printed material with 50% tonal range is likewise obtained. Differences in the size of the grid points of cyan, magenta and yellow can be determined by comparing these areas. These closed control methods also have the drawback that a separate colour bar has to be co-printed for this purpose. If this colour bar is not cut away, the appearance of the printed material is impaired. The co-printed colour bar is otherwise also used to monitor the positioning of the printed material on the paper web and/or to monitor the relative location (register) of the different colours, as described in the older European patent 0 850 763 of applicant. A drawback of both the open and closed control methods is that the printing press must be properly adjusted so as to ensure that a correctly printed colour bar also actually means that the total printed image is correct, since the colour bar is situated outside the normal printed image, and usually also differs greatly from the average printed image in respect of the colours used therein. Known from the above stated document US 5,744,635 is a method for monitoring printed material, wherein locating points are first defined in a digital file representing the original. Measuring points for the register, density, tonal range, colour and screening are then determined in this file and reference values are determined for the parameters. The ink supply is set in the first instance on the basis of these reference values. During printing records are made of the printed image, in which the locating points and measuring points are once again retrieved. The values of the different parameters in the record are compared to the reference values, and the printing presses are controlled on the basis of determined differences. US 6,024,018 further describes a method for monitoring printed material wherein a digital test record is first made of an approved proof. The RGB colour codes in the test record are converted into CYMK values and Regions of Interest (ROI) are defined in which reference values for the colours C, Y, M and K are determined. During printing records are made of the printed images in which the ROIs are once again retrieved. The conversion from RGB to CYMK once again takes place for each ROI, whereafter the values found are compared to the reference values, and in the case of differences the ink and/or water supply is adjusted. Although both these known methods have the advantage that the monitoring takes place in the printed material itself, and not in a colour bar outside the actual printed material, and they can moreover be performed in simple manner using computer-controlled equipment, they nevertheless have a number of drawbacks. Not every digital file is for instance automatically suitable to serve as original in a monitoring system as according to US 5,774,635, so that the results hereof are still not readily predictable. In addition, the proof which is taken as starting point in the system according to US 6,024,018 can of course already differ from the actual original. The invention therefore has for its object to provide a method for monitoring printed material as described above, wherein the above stated drawbacks do not occur. According to the invention this is achieved in a method as specified above in that the digital data file of the original is derived from a number of separate files for each of the colours (C,Y,M,K) to be printed. A very precise control of the quality of the printed material becomes possible by making use of the reference files for the individual colours. The digital data file of the original is preferably formed by converting to CIELab values the colours (C,Y,M,K) to be individually printed. Because CIELab values describe the colours in a manner which is closely related to human perception, an optimal control is thus achieved. A good comparison of the printed colours to the original becomes possible when the digitized inspection of the printed image comprises RGB colour codes which prior to comparison are converted into colour codes for each of the colours (C,Y,M,K) to be printed. For an optimum approximation to the original it is recommended that conversion takes place in accordance with a variable relation, wherein variations in the nature of the substrate, the printing ink, the printing press and the environment are taken into account. This can be embodied in efficient manner when use is made of lookup tables for the conversion. In order to be able to carry out monitoring properly, a pallet with a number of desired colours is preferably printed prior to manufacture of the printed material, the colours in this printed pallet are inspected and compared to the desired colours, and on the basis of this comparison a correction factor for the printing press is determined for each colour to be printed. A virtual proof of the image is advantageously shown on a display on the basis of the digital data file of the original. A client can thus already assess at an early stage how his/her printed material will finally look, whereby possible modifications can still be made at the last moment before actual printing is about to take place. The virtual proof can herein not only be displayed at the printer's, but can also be sent in digital form, for instance via internet, and can thus be displayed by the client on his/her own equipment. The digital data file can be pre-processed to correct differences during printing, in which case the reference values are preferably derived from the data file by undoing the pre-processing. Use is thus made for control purposes of the original in the form in which it was originally intended. The chosen parameters can comprise, in addition to the colours, the location of the images and/or the colour register of the printed material. Monitoring whether the printed material is lying in register can in this way be combined with monitoring of the colours. It is thus possible to dispense with co-printing of separate marks in the margin of the printed material for the purpose of register control, while automatic start-up can still take place, because when the printing press is started up it is already known what the printed image should look like. This method of controlling the colour register is moreover more accurate, since the colour register of the whole printed material is controlled, and not only that of the marks on the edge of the printed material. This register control integrated into the colour measurement also provides great advantages for monitoring the so-called fan-out, the deformation of the paper web as a result of it becoming wet, since the degree of fan-out depends on the degree of wetting, and thus varies greatly over the surface of the printed material. Colour photos, in which a lot of ink is used, will thus give a greater fan-out than text blocks with only a small quantity of black ink. A better control becomes possible by now monitoring the fan-out on the basis of inspections in the whole printed material, instead of only on the basis of a number of marks in the margin. An effective monitoring is achieved, when the reference values are determined in automatically chosen regions within the at least one original, and the inspection takes place in the corresponding regions in the printed image. It is thus not necessary to monitor the entire image, but only representative parts thereof, whereby monitoring takes less time and effort. These regions can be relatively small, to a minimum of for instance 0.1 x 0.1 mm2 when a determined colour shading occurs only in a detail of the image. When as is usual the colours for printing are built up of a number of basic colours, it is recommended that the regions are chosen such that each basic colour is present in at least one region. In this manner all colours can be monitored. A very good monitoring of the quality of the printed material is achieved when a number of inspections are performed for at least some of the colours, and a quality factor for the relevant colour(s) can be derived from the differences found. Differences can thus be clearly defined as a function of the tonal range. A rapid and simple control of the colours is obtained when the correction is performed by adjusting a quantity of ink applied to the substrate by the printing press. This quantity of ink applied to the substrate is advantageously then adjusted by adjusting a quantity of ink and/or water supplied to the printing press. Differences can thus be corrected quickly, without the underlying causes having to be known or having to be removed. Because some time will pass before a control operation has an effect on the inspected printed material, according to a particular aspect of the invention after a correction has been performed, a subsequent inspection and comparison is only carried out after a waiting time has elapsed. This prevents unnecessary control operations being carried out. According to another aspect of the invention, the magnitude of the differences is determined and subject hereto a choice is made for performing the corrections between at least a cruder and a finer adjustment. In the case of large differences the printed material can thus be quickly restored to within acceptable, relatively wide quality limits, while in the case of smaller differences the printed material can be adjusted somewhat more slowly but more accurately to an optimal quality. It is further recommended here that a warning signal be given when the magnitude of the differences exceeds a determined limit value. In this manner differences which cannot be corrected with the usual correction mechanisms, for instance because a printing plate is mounted incorrectly on one of the presses or the ink supply has become blocked, can be pinpointed immediately. In order to enable a rapid adjustment, according to yet another aspect of the invention inspection takes place while the printed material is still wet, wherein the detected colours for comparing are then corrected for colour changes during drying. The detected colours for comparing are preferably also corrected for ambient influences such as temperature and air humidity, so that the same quality of printed material is produced under all conditions. The printed material is preferably illuminated constantly during the inspection, so that ambient light cannot have an undesirable effect on the inspection. By setting the so-called light temperature the lighting can be adjusted as well as possible to daylight. The invention further relates to a system with which the above described method can be performed. Such a system is already known from the above mentioned older document US 5,744,635 which is provided with a device for determining, in at least one original for printing which is available in the form of a digital data file, reference values for one or more chosen parameters of the printed material, these parameters in any case comprising the colour of the printed material, a device for detecting the values of these parameters in a corresponding printed image, which detecting device is adapted to make and digitize a record of the printed image, and a device controlling the printing press connected to the determining device and the detecting device for comparing the detected values with the reference values, and for carrying out corrections when there is a difference in one or more parameters. The monitoring system according to the invention is now distinguished from this known system in that the determining device is adapted to read a separate file for each of the colours (C,Y,M,K) to be printed and to derive the digital data file of the original from these separate colour files. Preferred embodiments of the monitoring system according to the invention form the subject matter of the dependent claims 22 to 39. Finally, the invention further relates to a determining device, detecting device and comparing and correcting device which are intended for use in a monitoring system as described above. The device will now be elucidated on the basis of an embodiment, wherein reference is made to the annexed drawing, in which: Fig. 1 is a schematic representation of the main components of a monitoring system according to the invention in combination with a four-colour printing press, Fig. 2 is a schematic representation of the construction of one of the four printing presses, Fig. 3 is a flow diagram showing the most important steps of the method according to the invention for adjusting colours in printed material, Fig. 4 shows schematically how the reference values for the colours can be derived from the original digitized image, Fig. 5 shows a view of the paper web according to arrow V in fig. 1, in which the distribution of a number of regions over the image for printing can also be seen, Fig. 6 is a flow diagram showing the selection of the regions, Fig. 7 is a flow diagram which shows schematically the operation of the detecting device of the control system according to the invention, Fig. 8 is a flow diagram which shows schematically the operation of the comparing and correcting device of the monitoring system, Fig. 9 is a schematic flow diagram of the manner in which a suitable control is chosen by the comparing and correcting device, Fig. 10 shows an example of possible differences and the corrections applied therein by the comparing and correcting device, Fig. 11 shows schematically a possible correction via one of the ink keys of the printing press of fig. 2, as well as the effect thereof on adjacent parts of the printed material, and Fig. 12 shows an example of a curve representing the variation of grid point size as a function of the tonal range, Fig. 13 is a schematic representation of a part of the paper web having thereon an image with colours not in register, Fig. 14 is a schematic representation of a lighting unit to be applied in the monitoring system, and Fig. 15 is a flow diagram showing self-learning of the colours. A system 1 for monitoring and correcting printed material which is being produced by a printing installation 2, and which comprises one or more images I printed on a paper web 3, consists of a number of devices 4 to 6 co- acting with each other and with printing installation 2 (fig. 1) . These devices are respectively a device 4 whereby reference values for a number of parameters of the printed material, including in any case the colours for printing and in addition the location of illustrations and text, are determined in one or more originals for printing O, a device 5 which detects the values of these parameters, such as the colours in the relevant images I which are printed on paper web 3, and a device 6 for comparing the detected values with the reference values and for performing corrections in case of a difference in one or more values. The comparing and correcting device 6, which is connected to both the determining device 4 and the detecting device 5, controls printing installation 2. The operation of these different devices 4-6 and of printing installation 2 with all peripheral equipment is controlled as a whole by a general control system 7. This control system also provides the communication with operating personnel, and for instance displays status information and error messages on a display. Devices 4-6 of monitoring system 1 are connected to each other and to control system 7 via a network 21 developed by applicant. The comparing and correcting device 6 is connected to printing installation 2 via an network 22 forming part of this installation. In the shown embodiment the monitoring system according to the invention is applied in combination with an offset printing installation. The method and the system according to the invention could also be applied in other printing processes, such as flexo-printing or screen printing. Printing installation 2 is adapted to produce four- colour printed material, and comprises four printing presses 8. Each printing press 8 prints one of the basic colours C (cyan), M (magenta), Y (yellow) and K (black) on paper web 3, which is supplied from a paper roll 9. Paper web 3 is printed double-sided, and each printing press 8 therefore comprises on either side of paper web 3 a rubber cloth cylinder 10, a cylinder 11 engaging thereon which carries a photographic plate, an ink supply mechanism 12 connected to the photographic plate-cylinder 11 and a water supply mechanism 13 likewise connected to cylinder 11 (fig. 2). Both the ink supply 12 and water supply 13 are formed in the shown embodiment by a reservoir 14, 15 respectively filled with ink, water respectively, in which rotates a first roller 16, 17 respectively. This roller 16, 17 brushes over the surface of the ink respectively water, whereby this latter is entrained on roller 16, 17. It is otherwise also known in printing installations for newspapers to spray the water onto one of the rollers by means of spray nozzles, the so-called spray bar system. The quantity of ink entrained by roller 16 and transferred to a subsequent roller 18 is determined by so- called ink keys 19A, 19B, ... 19X (fig. 5), which are arranged distributed in longitudinal direction along roller 18. These ink keys 19A, 19B, ... 19N can be moved reciprocally transversely of roller 18 in the direction of the arrows, whereby the distance between the keys and the roller, and thereby the thickness of the ink layer on roller 18, can be varied. When newspapers are for instance printed on a so- called double wide/double size printing installation 2, in practice eight pages at a time are printed simultaneously on either side of paper web 3. This therefore means that on each photographic plate-cylinder 11 eight images 11-18, respectively I9-1-16 are formed. These images I are generally supplied by a publisher or editorial staff in digital form. At a printing company these digital files 23 are separated, usually in a preprocessing device 20, the so-called pre-press, into reference files per colour (C, Y, M, K), so that four colour files 25 are formed for each image I. These colour files 25 are one bit files - the colour is either present or not at a determined point in the image - with a very high resolution, and therefore form the basis of the monitoring according to the invention. These colour files 25 are usually already corrected on the basis of calibration graphs 24 (fig. 12) for the expected differences in printing presses 8. These calibration graphs or dot gain curves give for each value of the tonal range of the original, in this example plotted along the horizontal axis, the empirically determined value of the tonal range in the final printed image I. The corrected colour files can be expressed in a standard data format for printers, for instance TIFF/G4. In the preprocessing device 20a is determined how and where the images I must be printed on paper web 3, wherein the manner in which paper web 3 is folded and cut after printing is taken into account. The determining device 4 is adapted to already make, on the basis of the provided colour files 25, a "virtual" proof which can be shown on a display connected to the determining device 4, for instance the display of control system 7. The virtual proof can also be sent to an external computer, for instance via the internet. In the shown embodiment the determining device 4 is further adapted to determine from the digital files 25 with the print images I a number of regions, also referred to as ROIs (Regions Of Interest), which can be used to monitor the quality of the printed material. In this way it is not necessary to compare each image I fully with the original O, so that the number of inspections required and the associated computations remain limited. So as to be able to check the quality of the printed material against the original, this original 0 must first be derived from the digital files 25 supplied by preprocessing device 20. This takes place by undoing the corrections carried out in the preprocessing. On the basis of the same calibration graphs 24 used in the preprocessing the corrected data are restored as well as possible to the original inputted data. The thus obtained data files are subsequently compressed using applicant's own protocol, in order to allow the further processing to take place more rapidly. A search is then made in the compressed files for regions where the monitoring of the quality of the printed material can be carried out. Printed material generally consists of text blocks and illustrations, which are separated from each other by non-printed and therefore white areas or margins. In determining of the ROIs all parts of the original O separated by margins are successively assessed. For this purpose a search is made in a first region in the original (fig. 6, step 58) and the information from this region is read in (step 59) . A check is then made whether one or more of the colours C, Y and M are represented in this region (step 60) . If this is the case, the colour values of the located colours are then determined (step 61) in CIELab or a corresponding system for describing colours in a three- dimensional space. Such CIELab colour values describe the colours in a manner which shows a linear progression with the human perception of colours. The relevant data, together with the position of the region in the original 0, expressed in x,y-coordinates, are then stored in a ROI-file (step 62) . Also included in this ROI file is the location of the relevant image I in the printed material. This follows from the configuration 26 of printing presses 8, which is sent to determining device 4 shortly before the start of printing. A check is then made whether the entire original 0 has meanwhile been searched (step 63) . If this is not the case, the program returns to 58, and a search is made in a subsequent region within the original 0. When it is determined in step 60 that none of the colours C, Y or M are present in the region, this latter must then obviously be a black text area which in principle cannot serve as ROI. A jump directly to step 63 is then made. If in step 63 it is determined that the entire original has been searched, step 64 checks whether the colour black (K) is present in at least one of the located ROIs. If this is the case, the determination of the ROIs can then be stopped. If however there is still no ROI with the colour black therein, a search is then made for a black text region which must serve as ROI (step 65), and the colour values of the located colours are determined therein (step 66), whereafter these data with the position of the text part are again stored in a ROI file (step 67) . All colours are in any case hereby present in one of the ROIs, whereafter the determination can be ended. The format of the located regions or ROIs can also vary considerably, depending on the colours present therein. In order to isolate a specific colour shading in a ROI, it is thus possible to envisage the dimensions thereof amounting to for instance only 0.1 x 0.1 mm2. For each ink key 19 and each colour C, Y, M and K as many measurements as possible are carried out with varying tonal ranges. Although the greatest variations are detected particularly at values of the tonal range of around 50 percent, other values are of great importance for the quality of the printed material. For instance in the case of a very small tonal range of about 1 percent, any variation thus stands out sharply, although it might not catch the eye at a tonal range of 50 percent. In this manner the fullest possible dot gain curve is built up for each ink key and for each colour. Here the variations relative to the thus determined dot gain curves at different points in the printed image then finally form a measure for differences to be corrected in the printing press. The thus formed ROI files are passed from determining device 4 to the comparing and correcting device 6, where they are read and serve as reference against which the detected colours of the printed material are checked. The ROI-files are also sent to detecting device 5, so that this latter can perform inspections at the correct locations within the printed images I. A position recognition of each ROI from the original O per colour is herein also sent, so that on the basis thereof the inspection of the printed image I in x and y direction can be synchronized per colour with the ROI in the original 0. This is important in order to be able to also use the inspection for the control of the colour register, fan-out register and the like. In order to control the colour register, detecting device 5 first determines the position on the basis of the black colour K for a given ROI in the image I, for instance an illustration. This position on paper web 3 is measured in tenths of millimetres, in the x direction relative to the cutting position S of the repeating printed image (fig. 13), and in y direction relative to the mechanical centre CL of printing installation 2. The measurement values A, B are passed to the comparing and correcting device 6, where they are compared to the corresponding values in the original 0. The results of this comparison are used in turn to control printing installation 2 so as to hold the cutting position and the position of paper web 3 constant in the y direction. For each colour C, Y, M in the relevant ROI the distance is then measured relative to the colour black K in both the transverse direction L (laterally) and the peripheral direction C (circumferentially) of the cylinders, i.e. the y direction respectively the x direction of paper web 3. These measurement values are also fed to the comparing and correcting device 6, where they are again compared to the corresponding values of the original O. Possible differences, which therefore represent errors in the colour register, are used to control colour register correction motors of the separate printing presses 8, such that the colour register is once again restored. The algorithms used herein for searching the colours and determining the differences correspond for the greater part with those described in the older European patent 0 850 763 of applicant. In the shown embodiment the detecting device 5 comprises two scanners 27 placed on either side of the paper web. Each scanner 27 is here formed by a digital colour camera 28 with CCD-matrix, a lens 29 and a lighting unit 30. The CCD-matrix of the camera makes colour records in the RGB (Red, Green, Blue) format known from television. For financial reasons use is made of scanners 27 which have a field of view covering only a limited portion of the width of paper web 3, and which are each movable in transverse direction by means of a motorized traverse 31 (fig. 5) . The lighting units 30 are adapted to approximate to daylight as closely as possible by emitting a colour of light whose so-called light temperature corresponds to that of daylight. Each lighting unit 30 comprises for this purpose one or more light sources, for instance halogen lamps 68, which are each enclosed by a mirror 69 (fig. 14) . Lamp 68 and mirror 69 are accommodated in a dustproof housing 70, one side of which is formed by glass 71. Change in the light colour through time is avoided by excluding dust. In order to approximate as closely as possible to the light temperature of daylight, which amounts to about 55000K, mirror 69 is combined here with a filter 72 which is permeable to yellow light. In this way the light radiating directly from lamp 68 through glass 71 is thus combined with light reflected by mirror 69 from which the yellow component has been filtered. The detecting device 5 comprises a control part 32, for instance a computer, which drives scanners 27 to the correct position (y coordinate) on traverse 31 on the basis of the data from the ROI-files (fig. 7 step 34) . In order to determine the position in longitudinal direction of paper web 3 (x coordinate) use is made of a signal from a pulse generator 33, which is connected to one of the printing presses 8. After all, the angular position of printing press(es) 8 is at any moment directly linked to the position of the already printed images I in the longitudinal direction x of paper web 3. When scanner 27 is situated at the correct y coordinate and it is possible to infer from the signal from pulse generator 33 that the searched ROI is lying in the field of view of scanner 27, a record is made (step 35) . Computer 32 is further adapted to correct the records made by scanners 27 for differences which are the result of instability and imperfections in the equipment used. Examples of such differences are a non-homogenous lighting, variations in the intensity of the lighting, variations and inaccuracies in the CCDs, non-linearity of the measurement of the tonal range and the background colour of the paper. In addition, computer 32 is adapted to convert the measured RGB values into corresponding values in the CYMK format usual at printing companies (step 36) . Use is made for this purpose of lookup tables, comparable to CIELab colour tables, which include the corresponding values in these two formats. In order to compile these tables use is made of a neural network, which is initially "trained" by providing a large number of CYMK colours with matching RGB images. A "pallet" is first defined for this purpose in which all desired colours are included (fig. 15, block 73), whereafter this pallet is printed as well as possible (block 74) . The printed colours are then measured and linked in a three-dimensional space to a colour value in CIELab or a similar system (block 75) . Finally, the thus determined colour values are stored as standard setting or correction factor for the printing press (block 76) . These tables are calculated for each CCD on the basis of the starting point that the signal of the CCD is defined for each pixel by the quantity of incident light. This quantity of incident light can be described as:

u = t-Jf(x)-g(x)dx (1)

in which t is the time, f(x) the spectral distribution of the incident light and g(x) the spectral distribution of the CCD. This latter value is given by the manufacturer, while the spectral distribution of the incident light is determined by, among other factors, the ink colours, the paper type and the light source used. For the definition of colours for the purpose of determining the tables, applicant has developed his own method which provides a definition of each colour in three dimensions in a table. Herein the colour is first separated from the intensity (luminance) L, and the colour vector is then normalized and increased with a factor. This increment factor is necessary to be able to express colour differences in a number which corresponds with the inspection by the human eye. The relations used in the definition are:

L = (R + G + B) /3 R,G,B e [0, 1] (2)

a = (R/(R + G + B))1/Λ R,G,B e (0,1] (3)

g = (G/(R + G + B))1/'2 R,G,B e (0,1] (4)

From the RGB values produced by the CCD matrix the associated CYM values are derived, as stated, on the basis of the lookup table in computer 32 of detecting device 5. On the basis of the ratio between the values of C, Y and M on the one hand and K on the other as can be found in the reference files 25 for the relevant ROI for the derived CYM values, an associated K value is then determined. The values of C, Y, M and K thus finally found in the record of the ROI are sent from detecting device 5 to the comparing and correcting device 6 (step 37) . In the comparing and correcting device 6, formed here by a computer, the CYMK values received from detecting device 5 (fig. 7, step 38) are first corrected for inter alia the fact that detecting device 5 inspects the printed material while the ink is still wet (step 39) . In the shown embodiment the detecting device 5 is after all situated directly downstream of the final printing press 8 and upstream of a possibly present drying street in which the ink is dried. This arrangement ensures that the control can respond very quickly, but entails that the colours measured by detecting device 5 are not yet the definitive colours of the printed material. In the comparing and correcting device 6 there are therefore stored correction graphs which indicate the progression of each colour C, Y, M and K as a function of the drying time. On the basis of the known distance between each of the printing presses 8 and detecting device 5 on the one hand and the also known speed of paper web 3 on the other, it is possible to determine for each colour the elapsed time at the moment detecting device 5 is passed. At this time the necessary correction can then be read for the relevant colour in the correction graph. If desired, detecting device 5 of the monitoring system according to the invention can also be placed downstream of the drying street. In that case the final print image is detected and the use of correction graphs for drying can be dispensed with. In the comparing and correcting device 6 the colours are also corrected for differences resulting from variations in ambient influences, such as temperature and air humidity. Once the measured colours have been thus corrected, and the colours have in fact been obtained as they would be detected after drying of the printed material, the possible differences in the printed material can be determined (step 40) . For this purpose the value of the tonal range measured in the ROI for each of the colours C, Y, M and K is compared to the corresponding value of the tonal range in the reference file and converted to a relative or percentual difference P per colour as according to the relation:

Pcoiour = ( G - V) / V * 100 ( 5 )

in which G is the measured value and V the reference value of the tonal range for that colour. The average or equalized value DE of the dot gain of the four colours C, Y, M and K is then determined:

DE = (Pc + Pγ + PM + Pκ) / 4 (6)

whereafter the density DS is determined for each colour by making use of the following formula:

log10 (White / ink colour) (7)

This density is therefore an indication of the extent to which a colour is present above the average or, conversely, below the average in the printed image. The detected values of the equalized dot gain DE and the density DS for each colour and each ROI can be used per se as a basis for adjusting the printing presses 8 to a desired value, but it is also possible to average these values for a number of colours or a number of ROIs. It is then determined on the basis of the detected values of DE and DS whether the differences have to be corrected by adjusting the ink supply or by adjusting the water supply. To this end the detected differences relative to the dot gain curves (fig. 12) are collected at different locations in width direction of paper web 3 and converted to a quality factor (step 41) and a check is made as to whether all these differences correspond to each other (step 42). In case of differences on the dot gain curves which occur in ROIs over a part of paper web 3 not related to a single ink key, for instance in the case of a number of ink keys on one side of paper web 3 together, an adjustment of water supply 13 is the most appropriate way to restore a desired ink/water balance (step 43) . When on the other hand the differences on the dot gain curves can be related to the density differences and vary over the width of paper web 3 in an ink key range, a correction can take place via ink keys 19, which cover only a part of the width (step 44) . 5 The correction of water supply 13 as well as that of ink supply 12 has a number of different adjustments, one of which can be chosen depending on the magnitude of the detected difference. In the case of water supply 13 there is in the shown embodiment the choice between a normal 1.0 adjustment and a rougher but quicker adjustment, while in the case of ink supply 12 there is also provided a fine, somewhat slower adjustment. This is elucidated with reference to the control of ink supply 12. The absolute value of each detected difference DE 15 or DS is first determined (fig. 8, step 45), whereafter this is compared to a lower limit T0 (step 46). If the value is found to be smaller than this lower limit T0, there is no detectable difference and no correction is required. The program then returns to the start to read in 20 a subsequent difference and to determine the absolute value thereof. If the difference is greater than T0, it is then compared in a subsequent step 47 to a first threshold value T1, which determines a dead zone 55 (fig. 9) . When the 25 difference is smaller than T1, the measured value therefore lies within this dead zone 55 around the reference value REF, and the finer adjustment can in principle be chosen (step 48) . To this end one or more subsequent values of the difference are read in, made absolute and compared to the 30 first threshold value T1. If the absolute value of these differences is in each case smaller than the first threshold value, then there is indeed a good approximation of the reference value REF. In that case the finer adjustment is chosen, wherein the differences are averaged and, on the basis of this average difference, ink supply 12 is controlled such that the final result comes out in a fine zone 54 around the reference value REF. This control results slowly but surely in an accurate outcome. If on the other hand the difference is greater than the first threshold value, it is then compared in a following step 49 to a second threshold value T2, which determines an average zone 56. When the difference is smaller than this second threshold value T2, and the measured value thus falls outside the dead zone 55 but within the average zone 56, the normal adjustment is chosen (step 50) . One or more subsequent differences are here read in, made absolute and averaged with the preceding value(s) of the difference. The average is then compared again to the first threshold value T1. If this average is greater than the first threshold value, i.e. it is situated outside the dead zone 55, a correction is then carried out by adjusting the ink supply 12. Standard variations are thus corrected rapidly and with certainty. If the difference is greater than T2, i.e. it is situated outside the average zone 56, a relatively crude but rapid adjustment is chosen, wherein ink supply 12 is immediately adjusted without prior averaging with one or more subsequent measurements. Large errors are thus corrected immediately, and saleable printed material can be produced as quickly as possible. It is otherwise possible to envisage the difference being so great that it can no longer be remedied with the usual correction mechanisms. A situation which comes to mind here is that one of the printing plates is mounted wrongly on a printing press, or even on the wrong press, for instance an M-plate on the C-press. Another example is the ink supply becoming blocked. In these cases colours are printed at wholly incorrect positions, or are no longer printed at all. The monitoring system 1 responds to this type of great differences by generating a warning signal, which can be shown as a message on the display of the general control system 7, or which can take the form of activating an alarm light or bell. The operators can then stop printing installation 2 before large quantities of worthless printed material are produced. This enables major cost-saving, particularly during start-up. When ink supply 12 is adjusted by changing the setting of one of the ink keys 19, the effect hereof on the adjacent part of the printed material must be taken into account. Because an ink roller 18 is not compartmentalized, the ink layer defined by ink keys 19 flows out in width direction of ink roller 18, so that increasing the thickness of the ink layer in an ink zone - the central ink zone in the example in the lower half of fig. 10 - also results in an increase in the thickness of the layer in a part of the adjacent zones, as can be seen in the upper half of fig. 10. This effect can be compensated by either adjusting ink keys 19 in these relevant zones such that they dispense a slightly smaller layer thickness, or by modifying the setting of ink key 19 in the zone for correcting to a slightly lesser extent than would actually be optimal for the sought correction. In order to determine the effects on adjacent ink zones, use is made of a self-learning control system on the basis of tables which include for each ink key the percentual effects on two adjacent ink zones per side (therefore a total of four ink zones) . The actual effects are constantly measured and, on the basis of differences between the measured effects and the effects according to the tables, the table values can be slightly modified. Variations in these effects, for instance as a result of fluctuations in humidity and temperature, differences in the viscosity and/or wear, can hereby be compensated. Once a correction has been carried out, there is a wait time before a series of measurements is again carried out that are used for the relevant control. This is because the effects of corrections can only be detected with a delay and it is necessary to prevent the adjustment becoming unmanageable. After a correction signal has been sent to a controller 57 for ink keys 19 of one of the printing presses 8 and ink keys 19 have taken up their new position, the desired layer thickness of ink is dispensed onto ink roller 19. It will then however be some time before this ink is finally printed onto paper web 3 via plate cylinder 11 and rubber cloth cylinder 10. Some further time then passes before the relevant images I reach detecting device 5. The time T which elapses before the effect of a correction can be measured by detecting device 5 amounts to:

T = KD/MS + (MB/V)*100/lV + LB/V (8)

in which KD is the Δink value for correcting, MS the motor speed (in Δink/second), MB the distance (in metres) which the ink must travel through printing press 8 to reach paper web 3, V the speed of printing press 8 (in metres per second), IV the percentual transfer of new ink which causes the delay in the passage of ink in printing press 8, and LB the distance of printing press 8 from detecting device 5 (in metres). Expressed in the number of cylinder revolutions, the wait time WT becomes: WT = T / V * CD ( 9 )

in which CD is the diameter of rubber cloth cylinder 10 or the length of the repeating printed material. During the wait time the printed material is however monitored continuously in order to keep track of the quality development and optionally be able to carry out interim corrections in the case of sudden large differences. As stated, water supply 13 is controlled in similar manner to ink supply 12, be it that in the shown embodiment only two control levels are used herein, a normal and a rapid control. In the case of water supply 13 less account need be taken of the effects on adjacent zones, since water supply 13, at least in the shown system with a water reservoir 15 in which a water roller 17 rotates, is substantially constant over the whole width of the cylinders. The above described method and the associated control system make it possible to accurately monitor the quality of printed material and rapidly correct possible differences. Compared to the controls known heretofore, the invention provides a large number of advantages: Because the control is based on detecting and comparing the actual colours in the printed material instead of all kinds of printing engineering parameters which have an influence on colour stability, a very direct and accurate control is achieved. This is all the more so because for the description of the colours in both the original and the printed material use is made of CIELab or a similar system which describes colours in the way the human eye perceives them. Because the control corrects differences caused by both the printing machine and environmental influences, it provides a stable reproduction of the desired colours, irrespective of the location where the printed material is produced. This is a great advantage for publications printed in different printing firms at different locations, since the quality of the printed material can thus be kept constant. The so-called "colour management" process can hereby be considerably simplified. Because in the colour control according to the invention use is made of a direct comparison between the printed material and the original, which is derived from the supplied digital files, by undoing corrections which have been made thereto, variations in the printing installation, the quality of the paper and the inks do not have any significant effect. The regular calibration of the printing installation can hereby be dispensed with, which saves time and cost. The corrections carried out on the supplied digital files can also be used to determine an accurate presetting of the colours prior to printing. The quality of the printed material is hereby already very good immediately after start-up, and the use of correction graphs for the behaviour of the ink keys can be dispensed with. Because the colours and the colour register of the actual printed material are measured, a colour bar no longer need be co-printed, whereby paper, ink and preprocessing time are saved, while the appearance of the printed material moreover becomes more attractive. Measurement in the printed image itself also provides more and better information than a measurement in a relatively small colour bar outside the actual printed image. By making use of variable and intelligent conversion formulas instead of a fixed relation to convert the measured RGB values into CYMK values, variations in the inks used, the paper, the printing installation and the environment have no effect on the accuracy of the conversion. In addition, the control is very rapid, because the printed material is monitored immediately after leaving the final printing press. Saleable printed material is hereby already obtained shortly after start-up, while the colour consistency during the whole printing process is better than in systems where monitoring only takes place after drying of the printed material. The monitoring system can also be readily integrated into a printing installation owing to the chosen placing of the detecting device. This rapid control on the basis of an inspection immediately after the printing press is possible because use is made of correction graphs, with which the colour changes during drying of the inks is compensated. This compensation can be used in so-called coldset and heatset printing processes. The monitoring system can further be of simple design, because the detection of possible differences remains limited to relatively small regions of the printed material (ROIs) where the anticipated differences can be best detected. These regions can be found in efficient manner by the detecting device through a combination of a precise determination of the position in both longitudinal and transverse direction and the use of image recognition software. This enables start-up of the colour control while the printed material is not yet lying in register, whereby good printed material can again be produced very quickly. In addition, the manner in which the measurements of the detecting device are processed makes it possible to accurately derive both the density and the grid point size of the printed material from the measured values. This enables very good monitoring of the printed colours. Because the measured density and grid point size are combined in intelligent manner so as to determine the correction signals which are ultimately sent to the ink supply mechanism and the water supply mechanism, the final adjustment is moreover very accurate. Finally, the monitoring system according to the invention provides the option of also using the inspections of the colours and the comparison thereof to the reference image to control the colour register, the fan-out register, the cut-off register and the sidelay register. An integrated control of the total quality of the printed material is thus achieved, whereby a considerable simplification and saving can be realized compared to separate systems for the colour control and the register control. Because measurements are taken in the whole printed image, detailed distinction can herein be made between differences which result from deformation of the paper (fan-out) and differences which have other causes, and can be corrected by means of the colour register correction motors. When different printing plates are used, differences in the mutual position of the printing plates can also be measured and corrected in this manner. Although the invention is described above on the basis of an embodiment, it will be apparent to the skilled person that it can be varied in many ways within the scope of the following claims. Specifically, another system could be used instead of CIELab to describe colours in three dimensions, as for instance described above in respect of the conversion from RGB to CYMK, just as long as this approximates to the colour perception by the human eye. All new aspects described above are relevant per se for the invention, and could also be used in combination with other controls while retaining the advantages associated with the invention. The scope of the invention is defined solely by the appended claims.