CN106157300B

CN106157300B - Method for adapting an image examination by means of a computer

Info

Publication number: CN106157300B
Application number: CN201610465033.2A
Authority: CN
Inventors: F·绍尔特韦德尔
Original assignee: Heidelberger Druckmaschinen AG
Current assignee: Heidelberger Druckmaschinen AG
Priority date: 2015-04-20
Filing date: 2016-04-20
Publication date: 2021-07-27
Anticipated expiration: 2036-04-20
Also published as: DE102016204506A1; CN106157300A

Abstract

The invention relates to a method for adapting an image inspection by means of a computer, wherein a reference image is read in together with a global threshold value and a tolerance within the scope of an image inspection arrangement, wherein the image inspection detects a printed image of a printing process, indicates a deviation and a tolerance from the global threshold value on a display of a control computer, and determines a local accumulation of the deviation for a defective image. The method is characterized in that the user can change the threshold value and/or tolerance for a specific local region of the printed image during the printing process, and this changed local threshold value and/or tolerance is stored and used in the subsequent image examination.

Description

Method for adapting an image examination by means of a computer

Technical Field

The invention relates to a method for local threshold adaptation for an image inspection system.

The invention is in the technical field of test automation.

Background

Image inspection methods to be optimized have been used to date for inspecting finished printed products for defects that may occur within the scope of the execution of a print job. The printed image is scanned in by means of a digital camera, the data are sent to a computer, and a reference image generated from the data of the preliminary stage is compared at the computer. Alternatively, a defect-free sample of the printed image may also be used as the reference image. The integration thereof into an image examination is then referred to as read-in (Einlernen). Then, depending on the parameter settings of the comparison algorithm, a deviation between the re-digitized (readitalisier) printed image and the reference image is identified as a defect (Fehler) and indicated.

There are several possibilities for this parameter setting according to the prior art to date. The most common possibility is to manually adjust the setting parameters by the user. Here, the user analyzes a reference image created from data of the preliminary stage of the print job, and configures the image inspection method according to the analysis result and the user's personal experience. In addition to the threshold values, the following tolerance values (Toleranzwert) are additionally determined from the reference image: the measured value of the printed image to be inspected allows deviation from the threshold value by the tolerance value. A known extension of this method is, for example, to abstract the adjustment setting parameters to different sensitivity levels. This method has the advantage that even inexperienced users can carry out the parameter setting of the examination method, since the user no longer has to directly intervene in the parameters. Conversely, the disadvantage is that the flexibility of this method is insufficient compared with a completely free manual adjustment of the setting. Furthermore, a general disadvantage of all known manual methods relates to the dependence on a human user, i.e. the analysis of the user and thus the setting of the adjustment of the image-checking method may be erroneous. The less experienced the user, the more complicated the selection of the adjustment setting parameters, and the higher the probability of error. Furthermore, parameter adaptation is not usually performed on every new print job for time and cost reasons. Another disadvantage is that the configured tolerance is a global (global) value, the measured value allowing deviations from the reference image by said tolerance. That is, the maximum allowable deviation is equally large for all thresholds within the same sensitivity level. The problem that arises is that local (lokal) deviations of the reference image, which are not at all problematic for the user himself (for example due to the particular structure of these local deviations), are nevertheless marked as defects by the image inspection system, since these local deviations violate the maximum tolerance limit. The printed image often changes during the course of a print job (e.g., due to wear and other effects of the printer). This necessitates additional adaptations to tolerances during the print job.

The prior art to date has known solutions to this problem, for example, by covering the respective regions with a mask (Maske) and ignoring them by the image examination system (auszuSparen), which, however, causes gaps in the image examination at these points. Furthermore, the printed image can be divided into different inspection regions and then examined with the aid of different tolerance values. However, this division has disadvantages during the first configuration of the image examination: this division is only meaningful if it is immediately known which area of the printed image is problematic. Furthermore, such a division is mostly performed manually by the user and in practice is rarely performed manually by the user for the reasons of the required expenditure.

It is therefore known from DE 102014004555 a1 to increase the maximum tolerance for non-uniform image regions with multiple edges and to reduce the maximum tolerance for uniform image regions with few edges. However, this change is valid for all the requested inhomogeneous or homogeneous regions, i.e. in addition a global adjustment setting is also made. Furthermore, the global adjustment setting is also performed during the first configuration of the image inspection system.

Another possibility consists in re-reading in the complete reference image. Thereby changing the changed area to a new standard to be corrected.

The prior art does not allow for local adaptation of the maximum tolerance during running an image examination (that is to say for determined local image regions). By means of this local adaptation, the determined image regions and objects selected by the user can be marked with a maximum tolerance that differs from the remaining part of the printed image, so that these image regions and objects are adaptively verified by the image inspection system.

Disclosure of Invention

It is therefore an object of the present invention to provide a method for adapting tolerances of an image inspection system, which additionally adapts the tolerance values of the image inspection system for selected local image regions.

This object is achieved according to the invention in that a method for adapting an image inspection by means of a computer is proposed, wherein, within the framework of the image inspection arrangement, a read-in with a global threshold value and a tolerance is carried out, the image inspection system inspects the printed image of the printing process and displays deviations from the global threshold value and the tolerance on a display of a control computer, wherein the local accumulations of the deviations are summarized into a defect image. Characterised in that the user is able to change the threshold value and/or tolerance for a determined local region of the printed image during the printing process and these changed local threshold values and/or tolerances are stored and used in a subsequent image inspection.

Within the framework of the method, therefore, the image inspection system is configured as hitherto known from the prior art. However, if the system identifies a defect and displays it on the display of the control computer, the user can select the area of the image surrounding the identified defect and determine the size of that area and then determine the changed tolerance value for that area. The advantage of this method is that it is not necessary to read in a completely new reference image with adaptation tolerances, and the user can also react to changes in the printed image by adapting the maximum allowable tolerances. The change is also stored and thus made available for subsequent image examination.

Advantageous and further preferred developments of the method result from the corresponding preferred embodiments and from the description with reference to the corresponding figures.

In this case, a preferred development of the method according to the invention provides that a defined local region of the printed image is read in as a new local reference. If the changed tolerance is not sufficient, the selected local region can also be read in again as a new local reference. That is, as is not known in the prior art, the current page with the changed print image is read back as a new global reference, but only the local area selected by the user. This may be required, for example, if the individual printed objects within the printing area are geometrically offset. For example, if a barcode is shifted by two pixels in the printed image, the inspection area in the reference image that may have been printed cannot now be printed and vice versa. In this case, it is no longer helpful to change the maximum tolerance limit, but only to read in the shifted region again.

In this case, a preferred refinement of the method according to the invention provides that the threshold value and/or the tolerance in the determined local image region is increased.

In most cases, the maximum allowable tolerance needs to be increased. First, the local change, although exceeding the original tolerance, is evaluated as acceptable to the user.

In this case, a preferred development of the method according to the invention provides that the stored changed local threshold values and/or tolerances are used for repeated print jobs of the same print image.

If the same print job is repeated, it is recommended to use the changed local tolerance value again.

In this case, a preferred development of the method according to the invention provides that the stored changed local threshold values and/or tolerances are valid for the defect image.

Often, not only one defect in an area is identified, but a plurality of deviations directly adjacent to one another. In this case, the image inspection system collectively indicates these defects as one defect image. That is, the defective image portion on the display is extended to all directly adjacent defects, thereby indicating the entire defective area.

In this case, a preferred development of the method according to the invention provides that the defect image is associated with an image object of the printed image.

If a defect image having a plurality of adjacent defects is identified, the defect image is almost always associated with a certain image object in the printed image. The image inspection system identifies this defective image object and shows it completely. The user can then change the tolerance value for this image object or read in the local region again within the framework of the method according to the invention.

In this case, a preferred refinement of the method according to the invention provides that selected image objects are classified and, for a subsequent job with image objects of the same classification, the respectively assigned local threshold and/or tolerance of the selected image object is applied to this image object of the same classification. If a defective image object is identified and marked by the user with a changed tolerance value, the image inspection system stores this object and, if possible, classifies it into a specific category of image objects. Examples for such categories are for instance bar codes or matrix codes, sceneries, etc. Thus, the same tolerance changes are applied for image objects having the same classification in the current print job. Even in future print jobs applying image objects with the same classification, the same changed threshold and/or tolerance can be applied for this image object. This can be performed automatically or can be configured by the user.

In this case, a preferred development of the method according to the invention provides that the image object of the printed image is composed of edges.

The main application cases for such image objects identified as defective, but actually considered as good cases, are the following ones: the object has a plurality of edges, such as a bar code.

The method and its functionally advantageous refinements will be described in further detail in accordance with at least one preferred embodiment with reference to the respective figures. In the figures, corresponding elements are denoted by the same reference numerals.

Drawings

The figures show:

FIG. 1: a system configuration of the image inspection system;

FIG. 2: examples of defective image objects;

FIG. 3: preferred schemes of the process according to the invention.

Detailed Description

Preferred embodiments are described below. The image inspection system 2 shown by way of example in fig. 1 comprises at least one image sensor 5 (typically a camera 5) integrated into the printing press 4. At least one camera 5 captures the printed image produced by the printing press 4 and transmits the data to the computer 3 for analysis. The computer 3 can be a separate computer itself (for example, one or more specialized image processing computers 14) or the same as the control computer 3 of the printing press 4. The flow of the method according to the invention is schematically shown in fig. 3. The user 1 configures the image inspection system 2 in such a way that the user 1 reads the reference image 13 into the image inspection system 2. That is, for image inspection, the global threshold value 8 is calculated from the reference image 13, and then the image data of the printed image detected by the camera 5 is compared with the global threshold value 8. At the same time, user 1 assigns a permissible maximum tolerance 12 to threshold 8 via image inspection system 2. The permissible maximum tolerance 12 gives how far the brightness values of the actually printed image are allowed to deviate from the threshold value 8 of the reference image 13. The user 1 can set the tolerance value 12 of the entire printed image to a constant valueThis also corresponds to the general practice in practice. However, it is also possible to assign different tolerance values 12 to different regions of the printed image. In this way, for example, image regions with a uniform (homogen) surface can be sharper than those with a very non-uniform (homogen) surface (e.g., comprising a plurality of edges)

Verify and thus obtain a smaller tolerance value. These non-uniform surfaces must then accordingly achieve a high tolerance limit, since there is a high probability that they will deviate from the threshold value 8. With the aid of these threshold values 8 and tolerances 12 thus established, an image check is then carried out during the running printing process. If the image-checking system 2 finds a deviation 9 that exceeds the tolerance limit, it is displayed on a display 11 of the control computer 3 of the printing press 4. But if the user 1 is now able to accept the found defective image area 9, the user 1 is able to select the defective area or to select the found defective image object 7 (for example by controlling the mouse and keyboard of the computer 4 or by means of a touch screen) and assign this area 7 an increased maximum tolerance 10. For these defective areas or defective image objects 7, the control computer 3 of the printing press 4 stores the tolerance values 10 changed for the selected areas for running the image examination. Alternatively or additionally, the changed region can also be read in as a new local reference. This is particularly advantageous, in particular, when simply raising the tolerance limit 12 is not sufficient. Thus, this is shown in fig. 2: the image object 6 moves slightly in one direction geometrically on the sheet during the printing process with relatively sharp edges (here the bar code). In contrast to the reference image 13, the black area is now whitened due to the movement of the barcode 7 and vice versa. The image inspection finds a considerable number of defects 9. However, this is an acceptable drawback because the bar code 7 itself is correct. Therefore, the relevant regions are locally read as new references, so that the image inspection does not classify these regions into defective regions. For the remaining printing processes and the image checking which continues to run therewith, the selected printing process is then checkedThe region applies the changed tolerance value 10 or applies a new local reference. Furthermore, the user 1 can make adjustment settings in terms of the arrangement of image inspections so that the local change 10 performed remains valid even in the case of repeated print jobs.

List of reference numerals

1 user

2 image inspection system

3 control computer

4 printing machine

5 image sensor

6 original image object

7 image object with permission bias

8 Global threshold

9 defects found by image inspection

10 changed local maximum tolerance value

11 display

12 global maximum tolerance value

13 reference image read in

14 image processing computer

Claims

1. A method for adapting image inspections by means of a computer, wherein within the framework of an image inspection arrangement a reference image with a global threshold value and a tolerance is read in, wherein the image inspections inspect a printed image of a printing process and show deviations from the global threshold value and the tolerance on a display of a control computer, wherein local accumulations of deviations are summarized into a defect image,

it is characterized in that the preparation method is characterized in that,

the user can change the threshold value and/or tolerance for a determined local region of the printed image during the printing process and store and apply the changed local threshold value and/or tolerance to further image inspections and

only the determined local regions of the printed image are read in as new local references, and

the stored changed local threshold and/or tolerance is valid for such a defect image: a plurality of deviations directly adjacent to each other are collectively displayed as a defect image.

2. Method according to claim 1, characterized in that the threshold and/or the tolerance in the determined local image area is increased.

3. A method according to claim 1 or 2, wherein the stored changed local threshold and/or tolerance is applied for repeated jobs having the same printed image.

4. The method according to claim 1 or 2, wherein the defect image is associated with an image object of the printed image.

5. The method according to claim 4, characterized in that selected image objects are classified and for subsequent jobs with image objects of the same classification, the respective assigned local threshold and/or tolerance of the selected image objects is applied for the image objects of the same classification.

6. The method of claim 4, wherein the image object of the printed image has edges.