This patent application is a national phase filing under section 371 of PCT/EP2020/058953, filed Mar. 30, 2020, which claims the priority of German patent application 102019109791.6, filed Apr. 12, 2019, each of which is incorporated herein by reference in its entirety.
Apparatus for inspecting printed images for a printing or finishing machine and method for validating inspection algorithms of an apparatus for inspecting printed images.
Such an apparatus for inspecting printed images is known from German Patent Application No. 102012101310 A1, for example. It is used for the inspection of labels, for example, which are arranged on a carrier and are moved at high speed through a recording region of the apparatus for inspecting printed images. During this process the printed image is checked for defects.
Generally, the problem with print inspection systems is that it must be ensured that the inspection algorithms, including parameterization, have a specified sensitivity. However, the principle of error detection is not error-tolerant, i.e. if the inspection algorithms do not indicate an error, this means either that there are no defects or that they cannot be found with the current settings. For conventional systems, therefore, a sample roll with artificially introduced defects must be provided, which is then used to validate the inspection settings. This process is very time-consuming and due to abrasion, the sample roll must be replaced regularly, which involves additional costs.
German Patent Application No. 10 2016 100 437 A1 further discloses an apparatus for inspecting printed images that has a second illumination unit, whereby both illumination units are of different types.
Embodiments provide an apparatus for inspecting printed images, in which the validation is simplified. A suitable method shall also be indicated.
Embodiments provide an apparatus of the type mentioned above, which is characterized by an evaluation device, which is adapted to process at least two partial areas of the multi-line section as one strip image each, and has a validation mode, in which at least two strip images of a test image are compared with each other to check whether deviations of the strip images are detected.
An advantage of the apparatus for inspecting printed images is that the validation can be carried out on the basis of a static, i.e. standing image. While conventional systems require a moving test roll, the device can, for example, use a stencil with a test image which is placed in the image recording region of the apparatus for inspecting printed images. Due to the use of a camera with a multi-line image area, it is possible to capture two or more strip images without the need to move the test image.
In advantageous embodiments, a control device is provided and set up to set the evaluation device into validation mode according to a predefined time schedule, based on sensor measurement values, when a web change occurs or a test image is detected.
The test image is designed in such a way that a correct image is shown in a first area and a deviating image with one or more defects in a second area. During validation it is checked whether the defects built into the second area are correctly detected.
Regarding a method, the objective is solved by a method for validation of the inspection algorithms of an apparatus for inspecting printed images for a printing or finishing machine for continuously moving printed products based on matrix cameras with the steps
The invention is explained in more detail below using figures.
As described at the beginning, the quality of the printed images must be checked on printing or finishing machines. This can be done semi or fully automatically with the apparatus). For the image acquisition itself, the image acquisition device has a camera 7 with a CMOS camera chip. This enables the definition of a variable image format. With a device, the camera 7 captures at least one strip with a number of lines N, whereby the number of lines N is selected relatively small. As can be seen from
In this example, the area optically detected or detectable by the camera is larger than the strip with the number of lines N, but the image captured by the camera 7 is further processed in such a way that only the strip with the number of lines N is processed. In this example, the number of lines N is 70, whereby a line height is typically 0.1 to 0.2 mm. The height of the captured image strip is thus 7 to 14 mm. For printed products that are less problematic with regard to reflections, a greater height could be selected, for example 30 mm corresponding to a line number of 150 to 300, depending on the height of a line. For very difficult materials such as holograms, the number of lines could also be reduced, for example 50, resulting in a strip width of 5 to 10 mm.
The camera 7 is connected to an evaluation device 8, which assembles the strip-shaped images captured by the camera 7 and provides an overall image of a larger area. The evaluation device 8 also uses inspection algorithms to identify defects in the printed products. Among other things, this is done by comparing the captured images with stored reference images. When executing the inspection algorithms, setting parameters are used to influence, for example, the sensitivity of defect detection. In order to guarantee a constant level of defect detection, it is necessary to validate the inspection algorithms. This involves checking whether defects are detected with the desired accuracy and, if necessary, adjusting the parameters in order to achieve the desired accuracy again.
If a validation of the inspection algorithms is to be performed, a stencil with a test image is placed in the recording region 5 of the image acquisition device 3. This stencil can be designed as shown in
During validation, the system checks whether the defects built into the second area are detected. If not, the parameters are re-adjusted until the error detection works as desired again. If no suitable parameterization can be found, it is possible that the unsufficient error detection is caused by e.g. contamination of a lens. The validation thus refers to internal influences such as the type of inspection algorithm and the choice of sensitivity parameters, as well as to external influences such as illumination, degree of soiling of the lens or any deflection mirrors. Under certain circumstances, such external influences can also be derived from the defect image. If, for example, defects in the left part of the test pattern are correctly detected, but defects in the right part are not, an external influence in the optical part of the image capture device 3 is obvious.
In different embodiments, the time for the execution of a validation follows a predefined time scheme or is triggered by sensor measurement values. In addition, it is conceivable that a validation of the inspection algorithms will be carried out with each web change.
For operation, it is particularly easy if simply the stencil is placed in the recording area and a control device 6 automatically recognizes that it is a test image and then switches to the validation mode.
In an alternative embodiment, the control device 6 takes over the “control” and requests an operator of the apparatus 1 for inspecting printed images to insert a stencil with a test image.
As mentioned above, a matrix camera 7 is used for the apparatus 1 for inspecting printed images. “Matrix camera” means that a two-dimensional image with several image lines is captured. This is different from frequently used line scan cameras, which only take one image line and produce an overall image by moving the printed product under the camera, so that several images of the line scan camera simply have to be lined up.
In this point, apparatus for inspecting printed images with matrix cameras differ fundamentally from line scan camera systems. This is because the image sensor of a matrix camera 7 provides a large-area image 31, as shown in
During normal operation of a apparatus 1 for inspecting printed images, the strips are recorded and either evaluated individually or lined up together by image processing software to create an overall image. However, the image acquisition frequency is significantly lower than with a line scan camera, since each strip image already comprises several image lines.
In the apparatus 1 for inspecting printed images, the features of a matrix camera 7 are used, namely that two strip images can be recorded without moving a printed product or a test image under the camera 7. This is also shown in
In an advantageous embodiment, the two strip images 32 and 33 are captured in a single shot. In a modified embodiment, the strip images are acquired in separate images. It is possible to use adjacent areas in the recording region, as shown in
The validation procedure is explained below by means of a description of the methods.
In a first embodiment of a method, inspection algorithms are validated by placing in a first step a stencil in the recording region 5 of the apparatus 1 for inspecting printed images. The stencil has a first area 21 with a first pattern and a second area 22 with a second pattern, the first pattern having a defined deviation from the second pattern. Then, in a second step, a first multi-line strip image 32 is acquired, whereby at least a section of the first pattern is captured. In a third step, a second strip image 33 is recorded, whereby at least a section of the second pattern is detected. The two strip images can be acquired simultaneously, so that the second and third step coincide, or one after the other. Then, in a fourth step, the two strip images 32 and 33 are compared with each other and evaluated to determine whether the deviations between the first and the second pattern are detected.
If the first and second strip images 32 and 33 are acquired in a single shot, i.e. simultaneously, the position of the stencil for the first and second images is unchanged. Therefore, the selected area of the recording region of camera 7 must be shifted for the first and second image. If, on the other hand, the selected area of the camera is not to be changed, the stencil must be moved between a first and a time-staggered second image.
In order to facilitate positioning of the stencil, an advantageous embodiment provides a guide in the recording region 5 into which the stencil can be inserted for the validation procedure. This can be realized by a groove. A guide can also be used to facilitate moving of the stencil between a first and a second position by aligning the stencil at a first stop for the first acquisition and at a second stop for a second acquisition. This can be a left and right edge of a groove.
On the side of the evaluation device 6, an exact physical alignment of the stencil can be replaced by recognizing the position of a “somehow” inserted stencil and defining the sections of the recording region to be evaluated accordingly.
The apparatus for inspecting printed images is also suitable for surface inspection. In a surface inspection, for example, it is checked whether a product has holes. For this purpose, backlighting is used to check whether bright spots can be seen on the surface. A stencil 4o with a test pattern for a surface inspection is shown in
The invention is not limited to comparing exactly two strips. Test images with more than two strips could also be used, which are then evaluated.
Number | Date | Country | Kind |
---|---|---|---|
102019109791.6 | Apr 2019 | DE | national |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/EP2020/058953 | 3/30/2020 | WO |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2020/207841 | 10/15/2020 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
8270701 | Floeder | Sep 2012 | B2 |
8376516 | Viturro | Feb 2013 | B2 |
8611613 | Rauscher | Dec 2013 | B2 |
8947660 | Jaschul | Feb 2015 | B2 |
9719939 | Krebs | Aug 2017 | B2 |
9953435 | Fergen | Apr 2018 | B2 |
20110069166 | Rauscher | Mar 2011 | A1 |
20110170762 | Floeder | Jul 2011 | A1 |
20120013733 | Koltermann et al. | Jan 2012 | A1 |
20120105860 | Jaschul | May 2012 | A1 |
20150077538 | Krebs | Mar 2015 | A1 |
20150273816 | Schumann | Oct 2015 | A1 |
20170256047 | Schumann | Sep 2017 | A1 |
20190122352 | Krebs | Apr 2019 | A1 |
Number | Date | Country |
---|---|---|
102105888 | Jun 2011 | CN |
102428355 | Apr 2012 | CN |
102971619 | Mar 2013 | CN |
104246483 | Dec 2014 | CN |
106881956 | Jun 2017 | CN |
69300199 | Nov 1995 | DE |
102005010552 | Sep 2006 | DE |
102005059656 | Jun 2007 | DE |
102012101310 | Apr 2013 | DE |
102016204506 | Oct 2016 | DE |
102015114575 | Mar 2017 | DE |
102016203392 | Jun 2017 | DE |
102016100437 | Jul 2017 | DE |
2127878 | Dec 2009 | EP |
2407309 | Jan 2012 | EP |
2927868 | Oct 2015 | EP |
201400800 | Jan 2014 | TW |
2017121627 | Jul 2017 | WO |
Number | Date | Country | |
---|---|---|---|
20210252850 A1 | Aug 2021 | US |