CA2171891C

CA2171891C - Method for automatic evaluation by means of an otpo-electronic device

Info

Publication number: CA2171891C
Application number: CA002171891A
Authority: CA
Inventors: Luigi Stringa
Original assignee: De la Rue Giori SA
Current assignee: KBA Notasys SA
Priority date: 1995-03-30
Filing date: 1996-03-15
Publication date: 2007-05-08
Anticipated expiration: 2016-03-15
Also published as: JPH08290555A; RU2146622C1; ATE186016T1; CA2171891A1; IT1275707B1; UA34486C2; DE69604850T2; KR960035015A; AU4805296A; DE69604850D1; ITMI950637A1; CN1077042C; EP0734863B1; KR100421266B1; ITMI950637A0; CN1138693A; AU703121B2; EP0734863A1

Abstract

In order automatically to evaluate the printing quality, of an image by means of an optoelectronic device, the following procedure is adopted:
A plurality of reference images are prepared by known means and methods. Each reference image comprises an acceptable relative misalignment of various designs forming the image. The term different designs means designs printed during different printing phases.
During evaluation of the image for evaluation, comparison is made with the closest reference model.

Description

DE LA RUE GIORI S.A. LAUSANNE/SWITZERLAND
Method for automatic evaluation by means of an op,~
electronic device FIELD OF THE INVENTION
The present invention relates to a method for evaluation, using an optoelectronic device, of the printing quality of an image on paper comprising at least one design printed during a printing phase.
PRIOR ART
During the evaluation of printing quality on paper, in particular the printing of securities, use is made of electronic automatic inspection means comprising one or more black and white or color cameras for acquiring the images to be inspected. These images consist of matrices, usually rectangular, comprising numbers which represent the intensity of the light reflected or, in other words, the densimetric value of the pixels into which the image is subdivided. The number of pixels relating to an image depends on the resolution of the camera. In a monochrome (black and white) system, the image is described by a single matrix, whereas, in-polychrome ystems, the description consists of as many matrices as there are chromatic channels used. Conventionally, for descriptions of the RGB (red, green, blue) type, three chromatic channels are used.

The methods used for carrying out this type of automatic evaluation are based on the following schemes:
On the basis of a group of sheets considered as good, an acceptable printing quality model is constructed. Various techniques are used to construct this model. These include calculating, on the basis of the group of sheets considered as good, a standard mean image, that is to say an image which is described by a matrix in which each pixel is assigned the mean value which it has in the test sheet group.
Another method assigns two values to each pixel, one of which is the minimum value which it has reached in the test sheet group, and the other the maximum value. Two matrices are thus provided for each image, one with the minimum value and the other with the maximum value. Of course, in the case of a poly-chrome image, two matrices per color channel are obtained.
During the production of the images to be inspected, each pixel of the image to be inspected is compared with the pixel of the model thus obtained. If the difference exceeds a predetermined threshold value, or if it lies outside the minimum/maximum interval, the pixel is considered to have a printing defect. The number of defective pixels determines possible rejec-tion of the image, as a function of the quality which it is desired to obtain and which is determined beforehand.
During the production of certain types of valuable printed items, such as securities, bank notes, stamps, etc., the images are printed using various printing techniques such as offset, copper-plate, etc.
We will call these various types of printing printing phases. Thus, in a normal printing process, the paper first passes into a system for printing in the first phase, and a first design is applied, and it then passes into a second printing system for the second printing phase, making it possible to apply a second ~- 21~'i891 design to the paper. In this case, apart from the problem of printing quality, there is also the problem of coordination of the designs printed during the various phases. Indeed, deviations may exist between two images printed in this way, relating to drawings which are printed in different phases, even if only because of the deformation of the paper. These displacements, corresponding to a few pixels, may either be in the direction of displacement of the paper or in a perpendicular direction. In this case, it is no longer possible to extract a model which represents the desired printing quality by using the techniques mentioned above, because a single pixel may be assigned very different values as a function of alignments or correspondences between the printing phases.
In this case it has been proposed to construct a model for each printing phase. In order to do this, groups of sheets printed only with each of the printed phases are included in the test sheet group. With a procedure similar to the procedure mentioned above, a model is constructed for each printing phase. During the preparation phase of these models, the operator identifies the image portions which comprise only or essentially a lone impression phase.
During production, the relative misalignments between the printing phases are measured above all by using the pixels identified during the preparation of the models.
The models are subsequently combined, while taking into account the manner in which the various phases are successively printed on the sheets, in order to obtain a single reference model whose arrangement corresponds to the arrangement of the designs in the images for verification. Each image is then compared with the model thus produced. This method is complex and particularly expensive for the printer, because for each production batch it is necessary to print as many groups of sheets representing the desired printing quality as there are printing phases.

s The same inventor has proposed, in Canadian Patent Application 2,171,165, a method making it possible to determine, on the basis of a test sheet group, one model per design printed during a precise printing phase. The models thus obtained are subsequently combined to form a single reference model of the image for evaluation.
SUMMARY OF THE INVENTION
The object of the present invention is, on the one hand, to eliminate the problems of the prior art and, on the other hand, to propose an alternative to the aforementioned method of the same inventor. In the evaluation method according to the invention, a plurality of reference models are prepared, each corresponding to an acceptable misalignment, and the printed images are compared with the closest reference model or models.
The advantage of the present invention is its great flexibility with regard to the formation of reference models. Indeed, there are many ways of preparing the reference models with acceptable misalignments. For example, conventional methods may be used to choose the reference model of each of the drawings and subsequently to form reference images by combining said designs with relative misalignments which lie within acceptable values. The reference images are obtained either by calculation or by suitable printing. The model of each design can be obtained either by known means or by the method described in Canadian Patent Application 2,171,165.
According to a preferred variant of the invention, the reference models can be produced from test images, each image having an acceptable misalignment. According to another alternative embodiment, the reference models to be used are selected by the method making it possible to obtain the best correspondence between predetermined pixels of the reference models and their correspondants on the image for evaluation (best matching). The advantage of the present method is that it also makes it possible to reduce the r effects of sampling noise, which could invalidate the results when the misalignment between the optoelectronic device and the image for acquisition or for evaluation is less than the pixel dimensions of said device. Indeed, during such evaluation, the densimetric values acquired will be different if a black line is exactly centered on the sensor or if it is slightly offset.
With the aim of reducing the effects of such misalignment between the camera and the image for acquisition or evaluation, it is also possible to produce for each design forming the image reference models corresponding to possible deviations smaller than the pixel dimensions of the optoelectronic device. Thus, during evaluation, poor alignment between the camera and the image will not invalidate the evaluation result. Although this method requires greater use at first of the memory of the electronic system, there is no calculation to be carried out during the inspection phase and the method is therefore faster.
The invention therefore provides a method for automatic evaluation, using an optoelectronic device, of the printing quality of printed images on paper each comprising at least one design printed during a printing phase, wherein a plurality of reference models are prepared, each corresponding to an acceptable misalignment, wherein a misalignment of the printed images is measured, wherein a closest reference model corresponding to the measured misalignment is selected, and wherein the printed images are compared with the closest reference model.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention will be described in more detail by using two simple figures. Figures 1 and 2 represent two images for evaluation, which consist of two separate patterns whose relative positions are not the same; and Figure 3 schematically represents an installation making it possible to employ the method.

s DESCRIPTION OF THE EMBODIMENT
Figure 1 represents an image composed of a triangle Tl and a rectangle R1. Each of these designs is printed during a different printing phase, for example the triangle during offset printing and the rectangle during intaglio printing.
Figure 2 represents an image which also comprises a triangle T2 and a rectangle R2 which are also respectively printed during an offset and intaglio phase, the only difference between the two images being the fact that the relative positions of the rectangle and of the triangle are not the same. The misalignment has, of course, been exaggerated on this drawing. In principle, this misalignment during printing is only a few pixels.
According to the method of the present invention, evaluation of the printing quality and, in particular, evaluation relating to the misalignment between the two drawings forming the image, will be carried out by comparing each of the images represented in the two figures with reference models which consist of triangles and rectangles whose relative misalignments are within acceptable limits which have been predetermined.
As shown in Figure 3, a camera 2 inspects sheets 1 passing in front of it, and the acquired image enters a device 3 which makes it possible to measure the value of the misalignment. This misalignment value is fed to a memory in which all the models with acceptable misalignments have been indicated. This makes it possible to choose the correct model, which is then fed to a device 5, namely a comparator.
The printing quality is thus evaluated by using the models selected from the memory 4 with the image acquired by the camera 2.
It is obvious that this device requires a large memory, since all the models with acceptable misalignments must be stored in this memory, but on the other hand there is no need to make calculations during the evaluation, since it is sufficient to identify the model corresponding to the . CA 02171891 2006-10-18 s misalignment found by the device 3 in order to proceed with quality evaluation.
Production of these reference models is carried out as a function of the printing quality. Thus, when the printing tolerances (acceptable misalignment) have been fixed, the reference models are produced either by calculation or by starting with a suitable number of test sheets whose images are acquired into the memory of an electronic system. Upon processing of an image such as the one in Figure 1 or 2, this image is compared with the closest reference model and, if the differences are zero or within a predetermined range (tolerance), the image is accepted, otherwise it is rejected.
The reference image can be produced in other ways. For example, a reference model is produced for each design either by known methods or by the method developed by the invention forming the subject matter of the aforementioned patent application, and the image is recomposed with various relative alignments within defined tolerances.
It is clear that the number of reference models used determines the precision of the system. One method for selecting the models to be used during the quality evaluation is the technique making it possible to put certain predetermined pixels in correspondence between the reference models and the image for evaluation. In the case of the image presented in the two figures, it might be possible to use, for example, the lower vertex of the triangle and the far upper corner of the rectangle, referred to as A1, A2, and B1, B2 respectively. Thus, during evaluation, an attempt would be made to make pixels A1 and B1 or A2 and B2 on the image for evaluation correspond with corresponding pixels on the reference models.
This method also makes it possible to correct the effects of sampling noise which are related to the relative position of the camera and of the image for acquisition or evaluation. Indeed, the alignment ~mism between the camera and the image for evaluation or acquisition may not always be perfect. Thus if, for example, a black line whose width corresponds to that of the pixel exactly faces the sensor, its acquired densimetric value is different from that of the same black line when slightly offset, with an offset smaller than the width of the pixel.
In order to reduce or eliminate these effects of the camera/image misalignment, the same method can be used, that is to say the production of a number of reference models, also taking into account this misalignment between the camera and the image, which misalignment is strictly limited to the dimensions of the pixel. In this case, it is possible to produce models having, for example, misalignments of ~ 0.5 pixel with a step of 0.1 pixel, this being, of course, both in the x and y directions. For the abovementioned case, 121 models are thus created and are stored in the memory, for example the device 4 described in Figure 3 used hereafter for selecting the appropriate models and evaluating the quality. Here again, no significant increase in workload is produced, except as regards the employed memory of the electronic system.
It is clear that the technique of multiple ' models for the reduction of the sampling noise effect, even in the case of multi-phase printing, is applicable since it is sufficient to produce a plurality of reference models corresponding, for each printing phase, to various alignments between the camera and the drawing.
_ g _

Claims

1. A method for automatic evaluation, using an optoelectronic device, of the printing quality of printed images on paper each comprising at least one design printed during a printing phase, wherein a plurality of reference models are prepared, each corresponding to an acceptable misalignment, wherein a misalignment of the printed images is measured, wherein a closest reference model corresponding to the measured misalignment is selected, and wherein the printed images are compared with the selected closest reference model.

2. The method as claimed in claim 1, wherein the reference models are produced from test images, each test image having an acceptable misalignment.

3. The method as claimed in claim 1, wherein each printed image comprises a plurality of designs printed during different printing phases and wherein the reference models are formed by producing one reference model for each design printed during a corresponding one of the different printing phases, and by subsequently combining the reference models of each design in order to form the reference models with different and acceptable misalignments.

4. The method as claimed in any one of claims 1 to 3, wherein the reference models to be used during evaluation are selected by searching for the best correspondence between a predetermined number of pixels within the reference models and corresponding pixels within the images to be evaluated.

5. The method as claimed in claim 1, for reducing sampling noise effects, wherein the misalignment is less than a pixel size of the optoelectronic device and wherein the misalignment is a relative misalignment between the optoelectronic device and the printed image to be evaluated.