EP0012723B1 - Process for mechanically assessing the print quality of a printed product and device for performing the same - Google Patents

Abstract

Specimens, the quality of whose print is to be examined, are scanned photoelectrically point-by-point and compared point-by-point with one or more originals. The resulting reflectance differences are processed in different correction stages and then subjected to a point-by-point threshold decision, an individual threshold value being used for each image point. The threshold values are produced by analysis of specimens which have acceptable deviations, the maximum positive and negative reflectance differences due to their deviations being used directly as the threshold values. The analysis is effected by reference to electronically simulated specimens, an original or originals and a specimen being electronically displaced relatively to one another and reflectances being electronically varied in order to simulate register deviations and shade or tone deviations.

Description

The invention relates to a method for the mechanical assessment of the print quality of a printed product by point-by-point comparison of the test specimen to be assessed with a template, with the formation of the difference values between the reflectance values of the individual pixels of the test specimen and the reflectance values of the pixels of the template and corresponding to the specimen pixels Processing and evaluation of the difference values obtained in this way according to certain criteria, the evaluation comprising a final threshold value decision, and a device for carrying it out.

Such a method is e.g. B. described in FR-A-2 196494 and in DE-A-2 620 611. As can also be seen from this literature reference, one of the difficulties with such an automatic assessment method is to distinguish tolerable errors from intolerable errors in order to avoid incorrect assessments of the test object. So according to the aforementioned DE-A-2 620 611 z. B. smaller remission value differences between the test specimen and the template are eliminated by means of a minimum threshold correction, so that these small errors do not even enter into the further evaluation. The definition of this minimum threshold is critical. So there are z. B. in banknotes zones in which even the smallest color deviations are perceived by the eye as an error, and on the other hand zones, e.g. B. in the watermark, in which even relatively large deviations are still easily tolerated. In this regard, DE-A-2 620 611 states that the minimum threshold should not be the same over the entire image area, but could also be chosen locally, for example in the area of a watermark. Although this procedure already gives very good results, i. H. a relatively low frequency of misjudgements, it has been shown that these measures are not always sufficient.

The invention is therefore based on the object of improving a method and a device of the type defined at the outset in such a way that it works more reliably and leads to fewer misjudgments of the test specimens, which is less complex with the same quality requirements and achieves the above objectives with as little effort as possible .

This object of the invention is achieved by the measures specified in the independent claims.

The reference printed products used are preferably those which are subject to the largest possible deviations which are still tolerable. The errors should be of various types (positioning errors, register errors, tinting errors) in order to be able to record the effects of all errors that occur in practice on the machine test.

The invention is explained in more detail below with reference to an exemplary embodiment of a device suitable for carrying out the method according to the invention, which is shown schematically in the drawing.

The device shown is identical to the device described in DE-A-2 620 767, DE-A-2 620 765 and DE-A-2 620 611, with the exception of a few additional stages which are yet to be explained. It comprises three photoelectric scanning devices (scanners) 1 - for point-by-point photoelectric scanning of the reflectance values of a test specimen and two sub-image templates (1, 2), a relative position determination stage 4 for determining the relative positions between the test specimen and the individual sub-image templates, two displacement stages 5 and 6 for taking into account or compensating for the relative positions mentioned (register deviations), a combination stage 7 for electronically combining the image contents of the two partial image templates, a subtracting stage 8 in which the differences in the reflectance values of corresponding pixels from the test object and combined templates are formed, a tint correction stage 9, a minimum threshold correction stage 10, an error evaluation stage 11 operating according to the error recovery method described in DE-A-2 620 611 and a threshold decision stage 12 which, depending on the result of a p point-by-point decision about the value of a “good” or “bad” signal.

In this respect, the device shown fully corresponds to that described in the cited documents. In addition, the device shown also comprises two variable correction stages 13 and 14 with a transmitter stage 15 for setting the desired correction course, a position transmitter stage 16, via which the displacement stages 5 and 6 can be controlled in the same way as via the relative position determination stage 4, but independently of this, an electronic one Switch 17, a fault image memory 18 comprising a plurality of partial memories, a maximum value detection stage 19 and two threshold value memories 20 and 21 for the positive and negative threshold values, on the basis of which the threshold value decision stage 12 makes its good / bad decision.

Instead of the three separate scanning devices 1-3, of course, only a single scanning device and two suitable memories could also be provided, the individual partial image templates first being scanned sequentially and the resulting sample values being entered into the respective memory in terms of images should be written. The same applies to the shift stages 5 and 6, of which only one would also have to be present in the case of sequential operation. These and other possible variations of the device are within the skill of any person skilled in the art and therefore require no further explanation. It is also a matter of course that the entire electronic part of the device, insofar as it is not purely analogue areas, is expediently not implemented in hardware, but rather by means of a suitably programmed electronic computer.

If it is a matter of simpler printed products which are only produced by means of a single printing process, e.g. B. only in rotogravure or offset printing, or for print products with several printing processes, but lower quality requirements, of course, a single template with the overall image content is sufficient. In this case, the device would be reduced by the corresponding number of scanning devices or memories and the combination level.

High quality printed matter, how ?. B. banknotes and other securities, are usually produced in several printing processes using different printing technologies (gravure, letterpress, offset printing). In this case, the use proposed in DE-A-2 620 767 allows a plurality of sub-templates, the image content of which only corresponds to the image content of the printed product generated with one of the different printing technologies, for a more precise examination.

An essential prerequisite for this type of test is that the mutual positions of the test object and the template are known with reference to some stationary coordinate system (usually the scanning pattern of the test object). In practice, it is almost impossible to position the templates and the test objects in the scanning device in such a way that the scanned dots also match the respective pixels on the test object and the template (s).

In the position determination device 4 described in detail in DE-A-2 620 765, two pairs of relative coordination .DELTA.x, .DELTA.y between the respective test object and the two templates are therefore determined for each pixel according to the two templates.

The directly determined or stored sample values of the two templates are then shifted in the shift stages 5 and 6 by the coordinates dx, dx assigned to them by conversion so that all the pixels of the two templates match those of the respective test object. How this is done in detail is described in detail in the aforementioned DE-A-2 620 767. The correction levels 13 and 14 are inactive during the normal inspection of the printed products, i. that is, they have no influence on the reflectance values.

The remission values of the three sub-templates that have been shifted or corrected in this way are then linked to one another in the combination stage 7 by simple multiplication and then result in the overall template, which is compared in stage 8 with the respective test item point by point. The reflectance value difference Δl; form a difference image of the test object compared to the composite template. These reflectance value differences Δl; are first subjected to a tint correction in stage 9, an average value being formed from the difference values of a certain surrounding area of each pixel and subtracted from the difference value of the respective pixel. This tint correction is intended to avoid incorrect assessments due to minor tint deviations of the test specimen.

The difference values corrected in this way then arrive at the minimum threshold correction. stage 10, in which all those tint-corrected difference values which do not exceed a predetermined minimum threshold are eliminated, so that they are no longer included in the further evaluation. More about tint and minimum threshold correction can be found in DE-A-2 620 611, in which the following Fehlerberg evaluation stage 11 is also described in detail. An essential feature of the Fehlerberg method is that the difference values of the individual picture elements are not considered in isolation, but always in connection with the difference values of the environment points, whereby the respective environment points are still given a distance-dependent weight.

The difference values processed in this way then ultimately lead to the decision “good” or “bad” in step 13 ′ by means of threshold value detection. The threshold values required for this - one positive and one negative value per pixel - are located in the threshold value memories 20 and 21. Their detection or formation is described below.

The method according to the invention is based on the fact that even “good”, that is, in a visual inspection for test subjects who are found to be good, do not exactly match the template (s), but instead lead to certain remission value differences Δl when comparing in stage 8. The size of these remission value differences, their sign and their distribution over the entire image area, of course, depend on what was still considered permissible during the visual inspection and what was not. Experience has shown that most image errors are caused by register errors between the individual prints, by position errors in the watermarks and by color fluctuations. Other sources of error are image distortion and positioning errors between the test object and the previous day (s). The deviations permitted for each type of error are defined. According to the invention, the effects that all these approved errors have on the reflectance value differences in each individual pixel are now examined and the threshold values relevant for the error decision are determined in such a way that test specimens whose deviations from the template are still within the permissible range are actually be rated as "good". This setting of the threshold values is, of course, very critical, since it is very difficult to draw the line between "good" test specimens, ie test specimens with tolerable errors, and "bad" test specimens, because the effects of the different types of error on the remission value differences are very different. So it can e.g. B. may well happen that a tolerable register error causes a greater remission difference than an intolerable error in the watermark position.

According to the invention, an analysis is now carried out of test specimens which are subject to all possible but still just at the limit of the tolerable errors and the maximum positive and maximum negative reflectance value difference resulting from all these errors is determined for each pixel. For this purpose, for each type of error or for each test specimen, an “error image” composed of the individual difference values in each pixel is generated and stored in a separate partial memory of the error image memory 18 via the corresponding switch 17. The maximum value selection device 19 then looks for the maximum positive and the maximum negative difference value for each pixel from the individual partial memories and stores them in images in the two threshold value memories 20 and 21. These stored maximum difference values are thus directly used as individual threshold values for the good / bad Decision used in level 12. (If necessary, the maximum difference values can also be increased by a certain safety margin using an additive constant.)

To carry out this fault analysis or threshold value determination in practice, a large number of test objects would first have to be checked visually and then examined on the device. According to a further essential aspect of the invention, the error analysis is considerably simplified by not examining actual test objects, but electronically simulating such test objects and examining the simulated test objects. The maximum tolerable errors can then be conveniently set, and a few simulated test objects are sufficient to record practically all contingencies.

The register errors and position deviations are simulated using the position encoder stage 16 and the shift stages 5 and 6 controlled by it. For this purpose, either an almost ideally good printed product or one with medium register errors etc. is clamped as a test object and the relative positions compared to the template or templates determined by means of the relative position determination level 4. The template (s) are then successively shifted by the maximum tolerable distance in the four directions of the scanning grid and the shifted template (s) is compared with the test object, which in this case actually has a template function. In order to repeat it, the templates are shifted, of course not objectively, but only consist of an assignment of the reflectance values to pixels shifted by one or more pixel distances or a distance-dependent inter or extrapolation of the remission values in the individual pixels. The reflectance value differences which arise during these successive image comparisons, the totalities of which each represent an error image of the simulated test specimens concerned, are then stored in the error image memory 18 and processed further as described.

The simulation of faulty test objects can of course also be carried out entirely without a real test object by electronically creating an ideal test object from the templates, storing it and then using it as a comparison standard.

The simulation of register deviations between the individual prints of the printed product takes place by shifting the two templates relative to one another, and the simulation of positioning errors by shifting them simultaneously compared to the real or artificially generated test sample. A combination of both shifts is of course also possible.

The best way to simulate position errors of the watermark is to use two templates, one of which contains no watermark and the other only contains the watermark.

The two correction stages 13 and 14 and the variation transmitter stage 15 controlling them are provided for the simulation of tinting errors caused by printing ink or paper. These correction levels calculate the measured reflectance values supplied to them. B. according to the linear relationship

in resultierende Remissionswerte I_R um. I_w bedeutet darin den Remissionswert für irgendein Referenzweiß. Die Umrechnung bzw. Korrektur der Remissionswerte kann sowohl für die Neutrairemission (Gesamthelligkeit) als auch für eine oder mehrere Farbremissionswerte erfolgen. Entsprechend simuliert sie im einen Fall positive oder negative Neutraldichteabweichungen und im anderen Fall entsprechende Farbabweichungen gegenüber dem Vergleichsstandard.

into resulting reflectance values I _R. I _w means the reflectance value for any reference white. The conversion or correction of the reflectance values can be carried out both for the neutral emission (overall brightness) and for one or more color emission values. Accordingly, it simulates positive or negative neutral density deviations in one case conditions and in the other case corresponding color deviations from the comparison standard.

It goes without saying that the entire quality inspection can be carried out both in one channel (black and white) and in multiple channels (e.g. three basic colors).

The factor a in the above conversion formula can be set via the variation encoder stage 15. When checking the actual test objects later, the factor a is of course zero, so that the remission values pass the correction levels unchanged.

The method described above for obtaining the decision threshold values can of course also be used for such printed products; only a single template is used to check them, and in this case even more easily, since the number of possible errors is also lower.

For lower demands, instead of a positive and a negative threshold value for each pixel, either only the positive or only the negative threshold values can be determined and then stored in a single threshold value memory. The error decision is then made on the basis of an absolute residual threshold value comparison.

In addition to or instead of the electronic simulation of certain printing errors, a mechanical or optical simulation by physically moving or rotating the test specimen and previous day (s) or by inserting filters etc. into the scanning beam path can also be used

In the method described above, the definitive decision on errors only takes place after a longer and relatively complex preparation of the remission differences in stages 9, 10 and 11. However, the principle according to the invention of the individual evaluation threshold for each individual pixel also allows the error decision to be made at an earlier stage, for example after the tint correction level 9 or already directly after the comparison level 8, in which case the following stages would of course be superfluous. In this case, of course, the fault patterns of the simulated test specimens would also have to be obtained at the appropriate points, i.e. after the tint correction or directly after the difference has been formed, and the threshold values have been derived from them. These simpler variants of the test procedure are of course somewhat less sensitive and precise, but they enable a considerable reduction in the computing effort in cases where the demands on quality are not as high.

If the error decision is made directly in the difference field after the comparison level, whereby a test object is rated as bad or faulty if the remission value difference in a pixel or a predetermined number of pixels exceeds or exceeds the individual, possibly increased by the safety margin, positive or negative threshold value ., it is expedient to low-pass filter the reflectance values during the scanning in order to avoid pronounced error peaks and to achieve a more rounded course of the difference values over the image area. Suitable methods of low-pass filtering are explained in detail in the aforementioned DE-A-2 620 767.

The principle according to the invention of individual decision thresholds for each individual pixel enables on the one hand a refinement of the previously known test methods and on the other hand, with "lower quality requirements, a considerable reduction in effort. For example, it is then no longer necessary to fully compensate for positions and register errors in the quality test, but instead it is sufficient to render the errors which occur in the case of simpler and therefore less precise register deviation compensation harmless by raising the error threshold at the critical image points.

The quality inspection method according to the invention has a further advantage in that the individual error thresholds can be kept very "up to date". So if z. For example, a new production lot is available, some "good" test items from this lot are examined and their fault patterns compared to the templates are formed. If these error images contain larger errors than the previous error images, the relevant threshold values are replaced by the difference values in the relevant places of the new error images.

As already mentioned, with the exception of stages 13-21, all stages of the device are explained in detail in the three cited references DE-A-2620611, DE-A-2 620 767 and DE-A-2 620 765. These literature references also explain general problems of photoelectric scanning in the mechanical quality inspection of printed products, as well as suitable methods and devices for this.

Claims (14)

1. A method for assessing by machine the quality of the print of a printed product by point-wise comparison of an original and the sample for assessment, in which method the differences are formed between the reflectance values of the individual image points of the sample determined by poInt-wise photoelectric scanning, and the reflectance values of the image points of the original corresponding to the image points of the sample, and the resulting differences are processed and evaluated in accordance with specific criteria, evaluation comprising a final threshold decision based on threshold values individually associated with the individual image points, characterised in that a positive threshold value and a negative threshold value are used for each image point, said threshold values being obtained by analysis of a number of real or simulated reference printed products having visually acceptable faults, and ,being given increasing values the greater the maximum positive and negative deviations between the relevant image points of the original and the reference printed products as determined during the machine-testing of these reference printed products directly before the threshold value decision.

2. A method according to claim 1, characterised in that the threshold values used for each image point are in each case directly the maximum positive and the maximum negative deviation between the relevant reference image points and the original image points.

3. A method according to claim 1, characterised in that the threshold values used are the respective maximum positive and maximum negative deviations increased by a constant amount.

4. A method according to any one of claims 1 to 3, characterised in that the reference printed products used are electronically simulated or real products having electronically simulated faults.

5. A method according to claim 4, characterised in that position and registration faults are simulated by electronic relative displacement between the sample and the original or originals.

6. A method according to claim 4, characterised in that tinting faults are simulated by linear correction or conversion of the reflectance values in one or more colour channels.

7. A method according to any one of the preceding claims, characterised in that for each image point the reflectance difference values of the image points surrounding it are algebraically added, with distance-dependent weighting, to the reflectance difference value associated with each image point, and the threshold decision is made by reference to these added values.

8. A method according to any one of claims 1 to 6, characterised in that the reflectance values obtained from the photoelectric scanning are subjected to low-pass filtering and the threshold value decision is taken directly by reference to the reflectance value differences obtained from the point-wise comparison between the original and the sample.

9. Apparatus for performing the method according to claim 1 comprising a scanning stage (3) for the point-wise photo-electric scanning of the sample under assessment, a comparison stage (8) which compares the reflectance values obtained by point-wise scanning of the sample with the corresponding reflectance values of an original and at the same time forms the difference between the reflectance values for each pair of corresponding image points on the sample and the original, and comprising an evaluation stage (12) in which the differences of each individual image point are subjected to threshold value comparison directly or after processing in a processing stage (9 -11), characterised in that an analyzer stage (18-21) is provided which is adapted to be activated for a preparation phase and which, for each individual image point, determines the maximum positive and the maximum negative difference occurring . in each case in corresponding image points during a number of scanning operations and stores these two maximum values or two values absolutely exceeding these maximum values by a constant amount, the said values being stored as positive and negative threshold values for the evaluation stage (12), and that the evaluation stage (12) during the test phase following the preparation phase compares positive differences with the positive threshold value and negative differences with the negative threshold value.

10. Apparatus according to claim 8, characterised in that the analyzer stage comprises a maximum value-selector stage (19) and a store (20, 21) for each of the two maximum values.

11. Apparatus according to claim 9 or 10, characterised in that a simulation stage (13-16) is provided by means of which specific faults on printed products, or printed products themselves, can be simulated during the preparation phase.

12. Apparatus according to claim 11, characterised in that the simulation stage comprises a position variation sensor (16) for simulating a relative displacement between a sample and one or more originals.

13. Apparatus according to claim 11 or 12, characterised in that the simulation stage comprises a vario-sensor (15) and at least one computing stage (13, 14) for simulating tint deviations.

Images