METHOD AND DEVICE FOR OPTICAL INSPECTION OF CONTAINERS

Description

[0001] The invention relates to a method and a device for the optical inspection of containers having the features of the preamble of claims 1 and 12 respectively.

[0002] Typically, such methods and devices are used to inspect containers for contamination and / or defects. For this purpose, the containers are transported by a conveyor as a container mass flow to an optical inspection unit, which may include, for example, a lighting device and a camera so that the containers can be inspected using transmitted and / or reflected light. However, any other optical arrangements with which the containers can be imaged are also conceivable. The optical inspection unit captures each container as image data, which is then evaluated by an image processing unit for contamination and / or defects on the respective container. If contamination and / or a defect is detected in a container, it is, for example, cleaned again or recycled.

[0003] For example, such methods and devices are used for side wall, bottom and / or fill level inspection of empty containers or containers already filled with a product.

[0004] In rare cases, contamination spots may be present in the beam path of the optical inspection unit. These spots are then detected by the image processing unit as contamination and / or defects on one or more flawless containers. These cause them to be classified as defective and then inadvertently removed from the container mass flow. This can be detected, for example, by an automatic monitoring unit or by an operator, which then initiates cleaning of the optical inspection unit.

[0005] The disadvantage of this approach is that such contamination of the optical assembly is only detected if it leads to multiple consecutive poor inspection results or if the false rejection rate is significantly increased. Consequently, even slight contamination that does not lead to consecutive poor inspection results is often not detected. This can then lead to sporadic false rejections that go undetected.

[0006] An inspection device with at least one camera for inspecting objects, in particular containers such as bottles, cans or the like, and with an image evaluation device is known from DE 20 2004 007783 A1.

[0007] DE 196 46 678 A1 discloses a method for testing the reliability of a testing device, in particular an empty bottle inspector.

[0008] US 6,035,072 discloses a method for mapping defects or contamination that dynamically affect an image capture device.

[0009] The object of the present invention is therefore to provide a method and a device for the optical inspection of containers with which it can be more reliably detected whether contamination points are present in the beam path of the optical inspection unit.

[0010] To solve the problem, the invention provides a method for the optical inspection of containers with the features of claim 1. Advantageous embodiments of the invention are mentioned in the subclaims.

[0011] Extensive investigations by the applicant have shown that the contamination spots in the beam path of the optical inspection unit are always imaged at similar locations in the image data of several different containers. By overlaying the first image data of the multiple containers, the image signal of the contamination spots is amplified in the overlay image. In contrast, the contamination and / or defects on the respective containers are attenuated, as they are usually imaged at different locations in the image data. Even when only a few image data are overlaid, the contamination spots can be identified, for example, as a darkened area in the overlay image.By evaluating the overlay image for the presence of contamination spots in the optical inspection unit's beam path, these can be detected with exceptional reliability, and appropriate measures can then be initiated. For example, an operator can be warned early on so that they can clean the optical inspection unit. This method can detect even slight contamination spots, allowing appropriate measures to be initiated very early and without necessarily interrupting regular operations. For example, cleaning can take place during a conversion of a container handling machine assigned to the optical inspection unit.

[0012] The optical inspection method can be used in a beverage processing plant. The method can be installed upstream or downstream of a container manufacturing process, cleaning process, filling and / or closing process. The method can be used in a full-bottle or empty-bottle inspection machine that includes the optical inspection unit. For example, the method can be used to inspect returned reusable containers.

[0013] The containers can be intended to hold beverages, food, hygiene products, pastes, chemical, biological and / or pharmaceutical products. The containers can be designed as bottles, in particular as plastic bottles or glass bottles. Plastic bottles can specifically be PET, PEN, HDPE or PP bottles. It is also conceivable for the containers to be preforms for the production of the plastic bottles. They can also be biodegradable containers or bottles whose main components are made from renewable raw materials, such as sugar cane, wheat or corn. The containers can be provided with a closure, for example with a crown cap, screw cap, tear-off cap or the like. The containers can also be presented as empties, preferably without a closure.

[0014] It is conceivable that the method could be used to inspect the sidewall, base, mouth, and / or contents of the containers. Contamination could include foreign bodies, product residue, label residue, and / or similar items. Defects could include damage to the containers, such as chipped glass. It is also conceivable that they could be faulty material, such as localized thinning and / or thickening of the material.

[0015] The containers can be transported by the conveyor to the optical inspection unit as a container mass flow, preferably as a single-lane container mass flow. The conveyor can comprise a carousel and / or a linear conveyor. For example, it is conceivable for the conveyor to comprise a conveyor belt on which the containers are transported upright into an inspection area of ​​the optical inspection unit. Also conceivable are receiving elements that hold one or more containers during transport. The container can also be transported held by lateral belts, for example, if the lighting illuminates the container base and the camera inspects the base through the container mouth.

[0016] The optical inspection unit can comprise an illumination device and a camera. In the illumination device, the light can be generated with at least one light source, for example with a light bulb, a fluorescent tube and / or with at least one LED. Preferably, the light can be generated with a matrix of LEDs and emitted in the direction of the light exit surface. The light exit surface can be larger than the camera view of the container. It is also conceivable for the light exit surface to illuminate only part of the camera view of the container. The light exit surface can emit the light partially or completely diffusely. Preferably, the light exit surface can comprise a diffusion plate with which the light from the at least one light source is diffusely scattered over a large area towards the camera.

[0017] The camera can capture a portion of a container, precisely one container, or multiple containers, and the light transmitted or reflected across them using a lens and an image sensor. The image sensor can be a CMOS or CCD sensor, for example. It is conceivable that the camera transmits the image data, in particular the first image data or the second image data, to the image processing unit via a data interface. It is conceivable that the light is generated by the illumination device, subsequently illuminated by the containers and / or reflected by them, and then captured by the camera.

[0018] "Image data" here can mean at least one camera image of one or more of the containers. For example, the containers can each be captured in the first and / or second image data from one or more image perspectives. The camera can capture the polarization property, the intensity property, the color property, and / or the phase property of the light for each pixel of the image data.

[0019] The image processing unit can process the image data, in particular the first and / or second image data, with a signal processor and / or with a CPU (Central Processing Unit) and / or GPU (Graphics Processing Unit) and / or with a TPU (Tensor Processing Unit) and / or with a VPU (Vision Processing Unit). It is also conceivable for the image processing unit to comprise a memory unit, one or more data interfaces, for example a network interface, a display unit, and / or an input unit. It is also conceivable for the image processing unit to evaluate the at least one camera image using image processing algorithms that are present as a computer program product in the memory unit.

[0020] Overlaying the first image data of multiple containers to create the overlay image can mean summing the first image data of multiple containers. Optionally, the first image data of multiple containers can each be weighted with a weighting factor. "Summing" here can mean summing the same pixels from the first image data of multiple containers.

[0021] "That the overlay image is evaluated for the presence of contamination spots in the beam path of the optical inspection unit" can mean here that a local change in the overlay image is detected by means of the image processing unit, for example a darkening due to the contamination spots.

[0022] Another container can be captured with the optical inspection unit as second image data, wherein the overlay image is multiplied by a first weighting factor and the second image data is multiplied by a second weighting factor and then overlaid to form another overlay image, and wherein the another overlay image is evaluated for the presence of contamination spots in the beam path of the optical inspection unit. Subsequently, the another overlay image can replace the (previous) overlay image, and the previously described method step can be performed iteratively with yet another container. This means that only the overlay image and the second image data of the most recently inspected container need be stored in a memory unit of the image processing unit in order to calculate the another overlay image.The first image data of the previously acquired multiple containers can thus be deleted and therefore do not consume any storage space. Therefore, the first image data of a large number of containers, for example, 10, 100, or even more sets of first image data, do not need to be stored in memory in order to continuously determine a more current overlay image. The second image data and the first image data can be structurally identical and differ only in that the multiple containers were each acquired as the first image data and the further container as the second image data. In other words, the first image data and the second image data can differ only in that they were acquired from different containers.

[0023] The value of the first weighting factor can be greater than the value of the second weighting factor. Because the overlay image is already overlaid from several containers, the overlay image can be weighted more heavily than the second image data of the additional container. It is conceivable that both the first weighting factor and the second weighting factor are positive, or that both the first weighting factor and the second weighting factor are negative. For example, the first weighting factor can be proportional to the number of multiple containers and the second weighting factor can be proportional to exactly one container (i.e. to the one additional container). This means that both the overlay image and the second image data are weighted according to the number of containers inspected for each of them.

[0024] The first weighting factor and the second weighting factor can be in a range between 0 and 1, with the first weighting factor and the second weighting factor adding up to 1. As a result, the further overlay image has a similar value range to the (previous) overlay image and the second image data.

[0025] In other words, the first weighting factor can be calculated from the number of multiple containers divided by the number of multiple containers increased by 1, while the second weighting factor can be calculated from 1 divided by the number of multiple containers increased by 1. This makes it particularly easy to determine the weighting factors.

[0026] The first image data, the overlay image, the second image data and / or the further overlay image can each be normalized image by image to the same value range, in particular to a value range from 0 to 255 or from 0 to 65535. This allows the evaluation of said image data or the overlay images to be carried out particularly reliably using the same evaluation parameters.

[0027] It is conceivable that another container is captured with the optical inspection unit as second image data, with the overlay image and / or the further overlay image being multiplied by a third weighting factor and subtracted from the second image data in order to calculate the contamination spots in the beam path of the optical inspection unit out of the second image data. This allows the second image data to be cleaned up in such a way that the contamination spots are only very slightly visible or no longer appear at all. Consequently, the contamination and / or defects can be detected particularly reliably, even though the contamination spots are already present in the beam path of the optical inspection unit.

[0028] If contamination spots are present in the beam path of the optical inspection unit in the overlay image or in the further overlay image, a threshold sensitivity for evaluating the contamination and / or defects on the containers can be adjusted, in particular reduced. This results in fewer false rejections of containers that actually have no contamination and / or defects. It is conceivable that the threshold sensitivity is only reduced in certain areas. For example, the threshold sensitivity can be reduced only in an image area in which the contamination spots are depicted in the first image data and / or the second image data. Other image areas can remain unaffected. Consequently, the detection of contamination and / or defects in the image areas without the contamination spots is even more reliable.

[0029] Defective containers can be removed from the container mass flow after the contamination and / or defects have been assessed. This allows the defective containers to be excluded from further processing and recycled, for example, before they are filled with a product.

[0030] Before acquisition with the optical inspection unit, test image data can be acquired without the container. The test image data is then overlaid to form a test overlay image, which is then evaluated for existing contamination spots in the optical inspection unit's beam path. This allows existing contamination spots to be identified before the actual container inspection, ensuring particularly reliable inspection of the first containers. The test image data can be overlaid to form the test overlay image, for example, by summing them together. "Summing" here can mean summing the same pixels from the test image data. The test image data can, for example, consist of 10–20 test images without containers.

[0031] During detection, the containers can be illuminated with a plurality of light sources of the lighting device, wherein at least one of the light sources is activated on the basis of the overlay image in order to indicate the contamination spots to an operator for subsequent cleaning, and wherein the remaining light sources of the lighting device are deactivated. This makes it particularly easy for the operator to identify and clean the contamination spots. It is conceivable for the plurality of light sources to be arranged in a grid-like manner, for example in a tetragonal or hexagonal grid, in particular with the light sources being designed as LEDs. Preferably, the contamination spots can be assigned to the light sources on the basis of the overlay image. For example, during calibration, the light sources can be activated individually and image data can be recorded with the camera for each one.This allows the light sources to be assigned to the corresponding pixels of the image data and thus also to the pixels of the overlay image.

[0032] Alternatively, the contamination points can be displayed graphically to the operator on a screen, in particular the overlay image with markings of the contamination points.

[0033] It is conceivable that the method for optical inspection includes a manual or automatic cleaning process for the optical inspection unit, in particular the camera and / or the lighting device. This allows the contamination points to be removed from the optical inspection unit and the inspection of the containers can therefore be carried out more reliably. During the manual or automatic cleaning process, the lighting device can be cleaned with a cleaning agent, for example by flushing with a cleaning fluid. Preferably, a warning message can be displayed to the operator when an intervention limit is reached, for example on a screen, which triggers the cleaning process. In addition, the conveyor and / or an associated container handling machine can be stopped when the intervention limit is reached. The term intervention limit can refer to a degree of contamination.

[0034] Furthermore, to solve the problem, the invention provides a device for the optical inspection of containers with the features of claim 12. Advantageous embodiments of the invention are mentioned in the subclaims.

[0035] Because the image processing unit is designed to overlay the first image data of the multiple containers to form the overlay image, the image signal of the contamination spots is amplified in the overlay image. In contrast, the contamination and / or defects on the respective containers are attenuated, as they are usually depicted at different locations in the image data. Even with the overlay of just a few image data, the contamination spots can be identified, for example, as a darkened area in the overlay image. Because the image processing unit is designed to evaluate the overlay image for the presence of the contamination spots in the beam path of the optical inspection unit, these spots can be detected particularly reliably, and suitable measures can then be triggered.For example, an operator can be warned early enough to clean the optical inspection unit. This process can detect even slight contamination, allowing appropriate measures to be initiated very early without necessarily interrupting regular operations. For example, cleaning can be performed during a conversion of a container handling machine assigned to the optical inspection unit.

[0036] The device for the optical inspection of containers, in particular according to one of claims 12-14, can be designed to carry out the method described above, in particular according to one of claims 1-11. The device can comprise the features described above with regard to the method individually or in any combination.

[0037] The device for optical inspection can be arranged in a beverage processing plant. The beverage processing plant can comprise container processing machines, in particular a container manufacturing machine, a rinser, a filler, a closer, a labeling machine, a direct printing machine, and / or a packaging machine. It is conceivable that the device for optical inspection is assigned to one of the aforementioned container processing machines. It is conceivable, for example, that containers processed or manufactured by the container processing machine are optically inspected using the device. Likewise, it can be used, for example, to optically inspect preforms for the container manufacturing machine, preferably in a blow molding machine. The conveyor can connect one of the aforementioned container processing machines to the device for optical inspection and / or be assigned to it.The device can be used for full or empty bottle inspection. For example, it is conceivable that the device is designed for the visual inspection of returned reusable containers.

[0038] The inspection unit can comprise a lighting device for illuminating the containers and / or a camera for capturing the containers as image data. Image data here can refer to the first or second image data. This allows the containers to be illuminated or captured across their entire surface, so that contamination and / or defects can be detected particularly easily and reliably using the image processing unit. The lighting device can be designed to emit light across a large area from an extended light exit surface, in particular in the direction of the camera. The optical inspection unit can be designed to inspect the container between the lighting device and the camera. For example, the conveyor can be arranged between the lighting device and the camera in order to inspect individual containers of the container mass flow. This allows the containers to be inspected using transmitted light.It is also conceivable that the lighting device is designed for incident light and that the light emitted from it is reflected from the containers towards the camera during inspection.

[0039] The device may include a discharge device for removing defective containers from the container mass flow after evaluating their contamination and / or defects. This makes it particularly easy to recycle defective containers.

[0040] Further features and advantages of the invention are explained in more detail below with reference to the exemplary embodiments illustrated in the figures. Figure 1 shows an embodiment of an apparatus for optical inspection of containers in a perspective view; Figure 2 shows an embodiment of an apparatus for optical inspection of containers in a flow chart, in particular for implementation with the apparatus of the Figure 1 ; and Figures 3A - 3C show examples of the first image data, the second image data, the overlay image and the further overlay image as schematic image representations.

[0041] In the Figure 1 An embodiment of a device 1 according to the invention for optical inspection of containers 2 is shown in more detail in a perspective view. The device is used in conjunction with the device shown in the Figure 2 shown, inventive embodiment of the method 100 for the optical inspection of containers 2 is described in more detail.

[0042] In the Figure 1It can be seen that the containers 2 are transported by the conveyor 3 in the direction R as a container mass flow. The conveyor 3 is designed merely as an example as a conveyor belt on which the containers 2 are transported upright. However, any other types of conveyors that allow visual accessibility of the containers 2 are also conceivable. Only a single container 2 of the container mass flow is shown here as an example. It is understood that several containers 2 are transported on the conveyor 3, in particular arranged sequentially one behind the other, forming the container mass flow.

[0043] The optical inspection unit 4 comprises the illumination device 42 for illuminating the containers 2 and the camera 41 for capturing the containers 2 as image data. The containers 2 are scanned between the illumination device 42 and the camera 41 and thus optically inspected. However, an additional or alternative illumination device with incident light is also conceivable.

[0044] The illumination device 42 comprises the flat light exit surface 42a, which is designed in the manner of a luminous disc and which emits the light essentially homogeneously and over a wide area. However, any other type of illumination device 42 suitable for the respective inspection task is also conceivable. For example, the illumination device 42 comprises a matrix of LEDs that backlight a frosted glass disc, whereby the light is diffusely emitted from the light exit surface 42a. Directed emission of the light is also conceivable.

[0045] Also visible is camera 41, which captures the containers 2 as image data, particularly as camera images. For this purpose, camera 41 comprises, for example, a CCD or CMOS sensor and a lens. Both area-scan and line-scan cameras are conceivable.

[0046] Also visible is the image processing unit 5, which is connected to the camera 41 via an image data line 6. For example, the image data can be transmitted from the camera 41 to the image processing unit 5 as digital signals. It is also conceivable for the image processing unit 5 to be integrated into the camera 41. The image processing unit 5 is designed to superimpose the first image data I1, I2 of a plurality of containers 2A, 2B to form an overlay image 1 and to evaluate the overlay image U1 for the presence of contamination spots V in a beam path of the optical inspection unit 4.

[0047] Furthermore, defect D can be seen on container 2, which is depicted together with container 2 in the image data. Additionally or alternatively, contamination on container 2 is conceivable. Defect D is captured in the image data as a locally darkened area on container 2.

[0048] In addition, the contamination spot V can be seen in the beam path of the optical inspection unit 4, which also appears as a slight darkening in the image data behind the container 2. The contamination spot V is located here only, for example, on the light exit plate 42a of the illumination device 42. However, it is also conceivable that the contamination spot V is located on the lens of the camera 41 or on another optical element. For example, on a protective screen or on a mirror in a hall of mirrors.

[0049] The device from the Figure 1 according to the optical inspection method 100, the container 2 is removed from the Figure 2 used as follows: First, in step 101, the containers 2 are transported as a container mass flow by the conveyor 3.

[0050] In step 102, they are each captured as first image data using the optical inspection unit 4. It is conceivable that the containers 2 are captured individually or in groups.

[0051] The first image data are then evaluated in step 103 by the image processing unit 5 for contamination and / or defects D on the respective container 2. For example, an image processing method is used in which local grayscale changes in the image data are detected and marked. A threshold value can then be used to determine whether the corresponding container 2 exhibits contamination or a defect D.

[0052] However, it is possible that the contamination point V in the beam path of the optical inspection unit 4 is incorrectly classified as contamination and / or as a defect D. This then leads to the unintentional misrejection of otherwise defect-free containers. Consequently, the presence of the contamination point V in the beam path of the optical inspection unit 4 should be detected as early as possible.

[0053] In the Figure 3A It is shown by way of example that in order to detect the contamination point V according to step 108, the first image data I1, I2 of several containers 2A, 2B are superimposed to form an overlay image U1.

[0054] In the first image data I1 of the first container 2A, it can be seen that the first container 2A has a first defect D1. Furthermore, the contamination spot V is depicted on the illumination device 42 as a locally slightly darkened image area.

[0055] Furthermore, the first image data I2 of the second container 2B shows that it has a contamination D2. Furthermore, the contamination point V on the illumination device 42 can also be seen in the first image data I2 as a locally slightly darkened area.

[0056] During the overlay in step 108, the first two image data items I1, I2 are first normalized to a value range of, for example, 0-255 and then superimposed to form the overlay image U1. Preferably, the first two image data items I1 and I2 are added during the overlay. The overlay image U1 can then also be normalized to a suitable value range, for example, also from 0 to 255. In the overlay image U1, it can be seen that the containers 2A, 2B overlap and the defect D1 or the contamination D2 appears to be diminished. In contrast, the contamination spot V is enhanced so that it is now easier to recognize.

[0057] Subsequently, the overlay image U1 is evaluated by the image processing unit 5 according to step 109 for the presence of contamination spots V. Because the contamination D2 and / or defects D1 in the overlay image U1 are attenuated compared to the contamination spot V, the contamination spot V can be detected particularly reliably in the optical beam path of the optical inspection unit 4. For example, an image processing method is used to detect the contamination spot V, which detects locally darkened areas in the overlay image U1.

[0058] In a further step 104, additional containers are captured with the optical inspection unit as second image data I3 and I4, respectively. To detect the contamination point V even more reliably, a further overlay image U2 is determined from the second image data I3 of a third container 2C according to step 110. A further defect D3 and the contamination point V can be seen on the container 2C in the beam path of the optical inspection unit 4.

[0059] This is in the Figure 3B to be seen: For this purpose, the overlay image U1 determined in step 108 is multiplied by a first weighting factor and the second image data I3 is multiplied by a second weighting factor and then overlaid to form the further overlay image U2.

[0060] For example, to determine the overlay image U1 from the Figure 3Athe two containers 2A, 2B are recorded. Accordingly, the first weighting factor is calculated as the number of containers 2A, 2B divided by the number of containers 2A, 2B increased by 1, i.e., in this case, 2 / 3. Furthermore, the second weighting factor is calculated as 1 divided by the number of containers 2A, 2B increased by 1, i.e., 1 / 3.

[0061] It is therefore particularly advantageous for both weighting factors to be positive, with the first weighting factor being greater than the second weighting factor, since the overlay image U1 already contains the information from several containers 2A, 2B or their image data. In contrast, only a single container 2C was captured with the second image data I3, which is now to be additionally included in the overlaid information of the further overlay image U2.

[0062] The additional overlay image U2 is in the Figure 3Bright. In this image, the image information from the second image data D3 is weighted less than that of the overlay image U1. It can also be seen that in the further overlay image U2, the contamination spot V is higher than in the overlay image U1 from the Figure 3A is further enhanced and can therefore be recognized even more reliably.

[0063] The further overlay image U2 can now replace the previous overlay image U1 and steps 110 - 111 can be repeated iteratively with the second image data of further containers 2, wherein in yet further overlay images the contamination point V in the beam path of the optical inspection unit 4 then increasingly prevails over the contamination and / or defects on the containers 2.

[0064] In step 111, the image processing unit 5 evaluates the additional overlay image U2 for the presence of contamination spots V in the beam path of the optical inspection unit 4. This occurs analogously to step 109. If a contamination spot is present, then, following decision 112 in step 113, the threshold sensitivity for evaluating the contamination D2 and / or defects D1, D3 on the containers 2 is first adjusted so that, for example, detection based on the second image data in step 106 is less sensitive. The threshold sensitivity can also be adjusted only in a partial area of ​​the image data I1-I4 in which the contamination spot V is imaged.

[0065] Furthermore, the Figure 3C It can be seen that in step 105, using the example of a fourth container 2D, the contamination point V is calculated out of the second image data I4 using the further overlay image U2.

[0066] For this purpose, according to step 104, a further container 2 is acquired by the optical inspection unit 4 as second image data I4. It can be seen that the container 2D has a contamination D4. It can also be seen that the contamination point V in the beam path of the optical inspection unit 4 is still imaged as slight, local darkening. Furthermore, the yet further overlay image U3 can be seen, which was determined from a larger number of containers 2 according to steps 110 and 111. For example, image data from 10, 100 or even more containers 2 can be overlaid for this purpose. For this purpose, the second image data of yet another container 2 are superimposed on the further overlay image U2 according to step 110 and the yet further overlay image U3 is calculated in this way. Consequently, only the previous overlay image U1, U2 needs to be kept in memory and not the first orsecond image data of a large number of containers 2.

[0067] You can see further Figure 3C that, according to step 105, the contamination point V is calculated from the second image data I4 using the further overlay image U3. For example, the further overlay image U3 is multiplied by a third weighting factor and then subtracted from the second image data I4. It can be seen that, as a result, the contamination point V is no longer visible in the image data I5, so that the contamination D4 (or a defect) can be detected particularly reliably by the image processing unit 5 in step 106. This occurs in step 106 corresponding to step 103.

[0068] If the contamination point V now exceeds a predetermined level of contamination, the optical inspection unit 4 is cleaned in step 114. This can be done using a manual or automatic cleaning method, for example, by rinsing the light exit plate 42a with a cleaning agent.

[0069] During detection, the containers 2 can be illuminated with a plurality of light sources of the illumination device 42 arranged in a grid, for example, with a grid of LEDs. Based on the overlay image U1, at least one of the light sources is activated to indicate the contamination locations V to an operator for subsequent cleaning, while the remaining light sources of the illumination device 42 are deactivated. This can occur, for example, between steps 113 and 114.

[0070] Alternatively, the contamination points V can be displayed graphically on a screen to the operator before step 114, in particular the overlay image U1 with markings of the contamination points.

[0071] In addition, the containers 2 identified as faulty in step 103 or 106 are removed from the container mass flow in step 107.

[0072] By overlaying the image data I1, I2, I3 of the containers 2A, 2B, 2C, the image signal of the contamination spot V is amplified in the overlay image U1 or in the further overlay image U2. In contrast, the contamination D2 and / or defects D1, D3 on the respective containers 2A, 2B, 2C are attenuated, as they are usually depicted at different locations in the image data I1, I2, I3. Even with the overlay of just a few image data I1, I2, the contamination spots V can be detected, for example, as a darkened area in the overlay image U1. By evaluating the overlay image U1 or the overlay image U2 for the presence of the contamination spots V in the beam path of the optical inspection unit 4, these spots can be detected particularly reliably, and suitable measures can then be initiated.For example, an operator can be warned early on so that they clean the optical inspection unit 4. A manual or automatic cleaning method for the device 1, in particular the camera 41 and / or the lighting device 42, is conceivable. In the automatic cleaning method, for example, the lighting device 42 can be rinsed with a cleaning fluid. Preferably, a warning message can be displayed to the operator when an intervention limit is reached, for example on a screen. Furthermore, the device 1 and / or a container treatment machine assigned to the device 1 can be stopped when the intervention limit is reached. With the device 1 and the method 100, even slight contamination spots V can be detected, so that suitable measures can be initiated particularly early and regular operation does not necessarily have to be interrupted for this purpose.For example, cleaning can take place during a conversion of a container handling machine assigned to the optical inspection unit 4.

[0073] It is conceivable that, before step 102, test image data without the containers 2 are acquired with the optical inspection unit 4, wherein the test image data are superimposed to form a test overlay image, which is then evaluated for already existing contamination points V in the beam path of the optical inspection unit 4.

[0074] It is understood that features mentioned in the previously described embodiments are not limited to these feature combinations, but are also possible individually or in any other combinations.

Claims

1. Method (100) for optically inspecting containers (2), wherein the containers (2) are transported (101) as a container mass flow by means of a carrier (3), wherein the containers (2) are each detected (102) by an optical inspection unit (4) as first image data (11, I2), and wherein the first image data (I1, I2) are evaluated (103) by an image processing unit (5) for contamination and / or defects (D) on the respective container (2), characterized in that the first image data (I1, I2) of a plurality of containers (2A, 2B) are superimposed (108) to form a superimposed image (U1), and the superimposed image (U1) is evaluated (109) for the presence of contamination sites (V) in a beam path of the optical inspection unit (4).

2. Method (100) according to claim 1, wherein a further container (2C) is detected (104) with the optical inspection unit (4) as second image data (I3, I4), wherein the superimposed image (U1) is multiplied by a first weighting factor and the second image data (I3, I4) are multiplied by a second weighting factor and then superimposed (110) to form a further superimposed image (U2), and wherein the further superimposed image (U2) is evaluated (111) for the presence of contamination spots (V) in the beam path of the optical inspection unit (4).

3. Method (100) according to claim 2, wherein the magnitude of the first weighting factor is greater than the magnitude of the second weighting factor.

4. Method (100) according to claim 2 or 3, wherein the first weighting factor and the second weighting factor lie in a range between 0 and 1, and wherein the first weighting factor and the second weighting factor add up to 1.

5. Method (100) according to one of claims 2 to 4, wherein the first weighting factor is calculated from the number of the multiple containers (2A, 2B) divided by the number of the multiple containers (2A, 2B) increased by one, and wherein the second weighting factor is calculated from one divided by the number of the multiple containers (2A, 2B) increased by one. (2A, 2B) increased by one.

6. Method (100) according to one of claims 1-5, wherein the first image data (I1, I2), the superimposed image (U1), the second image data (I3, I4) and / or the further superimposed image (U2) are each normalized image by image to the same value range, in particular to a value range from 0 to 255 or from 0 to 65535.

7. Method (100) according to one of claims 1-6, wherein a further container (2D) is captured with the optical inspection unit (4) as second image data (I3, I4), and wherein the superimposed image (U1) and / or the further superimposed image (U2) is multiplied by a third weighting factor and subtracted from the second image data (I3, I4) in order to calculate the contamination sites (V) in the beam path of the optical inspection unit (4) from the second image data (I3, I4).

8. Method (100) according to one of claims 1-7, wherein if contamination sites (V) are present in the beam path of the optical inspection unit (V) in the superimposed image (U1) or in the further superimposed image (U2) (112), a threshold sensitivity for evaluating the contamination and / or defects (D) on the containers (2) is adapted, in particular reduced (113).

9. Method (100) according to one of claims 1-8, wherein defective containers (2A, 2B, 2C) are ejected (107) from the container mass flow after evaluation of the contamination and / or defects (D).

10. Method (100) according to one of claims 1-9, wherein, prior to detecting the containers (2) with the optical inspection unit (4), test image data is detected without containers, and wherein the test image data is superimposed to form a test superimposed image, which is then evaluated for existing contamination sites (V) in the beam path of the optical inspection unit (4).

11. Method (100) according to one of claims 1-10, wherein the containers (2) are illuminated during detection by a plurality of light sources of an illumination device (42), and wherein, on the basis of the superimposed image (U1), at least one of the light sources is activated in order to indicate the contamination sites to an operator for subsequent cleaning, and wherein the remaining light sources of the illumination device (42) are deactivated.

12. Device (1) for the optical inspection of containers (2), in particular for carrying out the method (100) according to one of claims 1 to 11, with - a carrier (3) for transporting the containers (2) as a container mass flow, - an optical inspection unit (4) for detecting the containers (2) as first image data (I1, I2), - with an image processing unit (5) for evaluating the first image data (I1, I2) for contamination and / or defects (D) on the containers, characterized in that the image processing unit (5) is designed to superimpose the first image data (I1, I2) of several containers (2A, 2B) to form a superimposed image (U1) and to evaluate the superimposed image (U1) for the presence of contamination sites (V) in a beam path of the optical inspection unit (4).

13. Device (1) according to claim 12, wherein the optical inspection unit (4) comprises an illumination device (42) for illuminating the containers (2) and / or a camera (41) for detecting the containers (2) as image data (I1 - I4).

14. Device (1) according to claim 12 or 13, wherein the device (1) comprises an ejection device for ejecting defective containers (2A, 2B, 2C) from the container mass flow after the contamination and / or defects (D) have been evaluated.