Method for inspecting a container by means of position determination

Abstract

A device (1) for checking objects (10), in particular beverage containers (10), comprising: a transport device (2) for transporting the object (10) along a predefined transport path; at least three image recording devices (12, 14, 16, 18) positioned so that they record spatially resolved images of the transported object (10) from different directions, wherein the device (1) comprises a position detection device (20) for detecting at least one position of at least one section (10a) of the object (10). According to the invention, the position detection device (20) comprises: a first marking device (22) suitable for and intended for temporarily providing a marking (30) to at least one surface section (10a) of the object (10); and a first sensor device (12, 14, 16, 18) suitable for and intended for detecting the marking (30) temporarily located on the surface (10a).

Description

Technical Field

[0001] The invention relates to a device and a method for checking objects and in particular containers. Various methods and devices are known from the prior art. The invention relates in particular to checking labels which are applied to objects and in particular containers. Background Art

[0002] The purpose of such label inspections is to scan objects and in particular containers all-round and to check existing labels at different points, i.e., for example, the presence of the label, its position, its tilt, the reading of a barcode or also the reading of a best before date (MHD), etc. However, printed materials, etc., can also be checked, for example.

[0003] The orientation of the container with respect to its longitudinal direction is arbitrary. It is known from the prior art to scan such containers with the aid of a plurality of cameras. For example, a matrix camera can be used here. In the devices known from the prior art, preferably three, four or six cameras are used in a plane in a circular manner. In addition, there is at least one observation plane, but preferably also two or more planes.

[0004] It is also known from the prior art to combine the individual images taken of the container to form a panoramic image. It is also possible to correct the distortions resulting from the imaging from one perspective and to stitch them together in a suitable manner with the other corrected images to form a panoramic image. It is possible to carry out the correction with the aid of simple assumptions, although more complex processing methods are also possible, such as 3D reconstruction.

[0005] For this purpose, in the prior art, the intersection line between two adjacent photographs is determined in order to achieve seamless stitching in the panoramic image using various techniques.

[0006] It is thus possible to determine the position of an object or container at the edge in one or more camera images. Here, the applicant's internal prior art is known to use a process of black and white patterns in the recognition, or also a process of cross-striped backgrounds, which is also known from the applicant's internal prior art.

[0007] Furthermore, it is also conceivable to alternatively determine the position of the container by means of a camera, for example from above.

[0008] The known container shape and its exact position allow the spatial surface position to be determined precisely. The seam line can be determined precisely in the two partial images from the exact surface position, the position of the two adjacent cameras and the intrinsic and extrinsic imaging parameters of the two adjacent cameras.

[0009] However, if these values, ie the positions, are not exact, then an attempt can be made to find so-called feature points in the overlapping area of ​​the seam in the adjacent images and to align them. Based on these feature points, the seam line can in turn be determined.

[0010] Finally, the rectified partial images are stitched together at the seam lines, thus producing a seamless panoramic image overall.

[0011] However, various errors, such as brightness differences in adjacent photos, can result in a recognizable transition at the seam line. Overlapping transitions, where the weight of one photo decreases continuously while the weight of another photo increases, can reduce this abrupt transition.

[0012] However, disadvantages arise in the prior art if the container position cannot be determined precisely. The assumed surface spatial position is then no longer applicable. Another disadvantage arises if the container shape deviates from the assumed shape. Another disadvantage may arise if the shape of the container deviates from the assumed shape. The assumed surface spatial position is then also no longer applicable.

[0013] If the contour of the container can be determined exactly by means of a camera, then the position of the contour or contour line will never coincide with the position of the seam line due to the camera being set up.

[0014] The disadvantage of determining the seam position by feature points in the above manner is that it is not necessarily guaranteed that such feature points can be determined in both images and can be clearly associated. This can be the case, for example, if both cameras have a very wide stereo baseline. The image contents are also captured from very different angles.

[0015] In addition, small parts and large-area patterns usually have a large number of feature points, which are always the same. In this case, the correct association is not always guaranteed. It is also conceivable that these feature points are completely missing, that is, the feature points cannot be found because the wrong features were selected. In this case, expert knowledge and knowledge about the expected image content are required.

[0016] Therefore, the following disadvantages also occur in the prior art: discontinuities in the panoramic information occur at the seams. It is possible that information is doubled or that information is missing. This makes it very difficult to identify the content, i.e., to read a 1D code, read a 2D code, read a text, etc.

[0017] In the prior art, the check is usually indicated only as a plausibility check. This means, for example, that if complete information can be detected depending on the orientation of the container in the camera, then this information is recognized when it extends over two image segments. If this information is not clearly detected, the check is not performed or is only partially performed. In this way, in the prior art, it is still a long way from being possible to carry out a 100% check of every container and in any orientation. Summary of the invention

[0018] The object of the invention is therefore to eliminate influences on the position determination, the contour determination, and ultimately also sources of error when determining the actual container surface, for example, by silhouettes. This object is achieved according to the invention by the subject matter of the independent claims. Advantageous embodiments and improvements are the subject matter of the dependent claims.

[0019] The device according to the invention for inspecting objects, in particular beverage containers, has a transport device that transports the objects along a predefined transport path. The device also has at least three image recording devices that are positioned so that they record spatially resolved images of the transported objects from different directions, wherein the device also has a position detection device for detecting at least one position of at least one section of the object.

[0020] In this case, in particular, the position detection device is suitable for and is intended to detect a position of an object or a portion thereof relative to at least one of the image recording devices or preferably relative to a plurality of image recording devices.

[0021] In this case, it is particularly preferred that at least one image recording device is arranged in a stationary manner, and preferably a plurality of image recording devices are arranged in a stationary manner, and particularly preferably all image recording devices are arranged in a stationary manner.

[0022] According to the present invention, the position detection device has a first marking device, which is suitable for and designed to temporarily set a mark (especially a detectable mark, and especially a mark detectable by means of an optical mechanism) on at least one surface segment of an object; and a first sensor device, which is suitable for and designed to detect a mark temporarily located on the surface.

[0023] The marking temporarily on the surface is understood in particular to mean that the marking is not permanently retained (i.e., for example, embossed) on the object or container, but is only temporarily retained on the object or container, in particular for inspection purposes. Of course, the marking can also be located on the object during the time period of creating the image photograph.

[0024] Preferably, the sensor device is suitable and designed for optically detecting the marking. Preferably, the sensor device is suitable and designed for detecting the marking together with other areas of the container and in particular together with other surface sections of the container.

[0025] In the prior art, attempts are made to extract suitable information via theoretical assumptions or from the image content of the object. It is proposed within the scope of the present invention to dispense with these aids and to apply the information, in particular contactlessly and temporarily, to the surface and in particular to the expected joint areas.

[0026] This information can be localized by the identification itself or by additional measuring devices and preferably also the actual surface position of the container and in particular the surface position in the seam region can be determined.

[0027] Preferably, the actual surface position is used to determine seams in adjacent camera images.

[0028] Thereby, error sources occurring in the prior art when determining such a container surface should be determined.

[0029] Preferably, the object is a container, and in particular an equipped container, and in particular a labeled container. The container is particularly preferably a bottle, to which the label is applied. However, it can also be an object and in particular a container, on which a printed matter, in particular a direct printed matter, is arranged.

[0030] However, it is also possible to check cans or shrink packaging or the like. Thus, the surface can also be, for example, the surface of a can and in particular a printed surface. In addition, the container can also be a plastic container, such as a plastic cup.

[0031] In this context, the containers can be used in particular in the fields of the beverage industry, body care, health care and food industry.

[0032] In a preferred embodiment, (at least) three image recording devices are arranged substantially in the same plane. Particularly preferably, the image recording devices are arranged laterally beside the transport path and / or laterally beside the container to be inspected.

[0033] It is conceivable that the image recording device views the container in a horizontal direction. However, it is also conceivable that the image recording device views the container from an inclined direction, for example, diagonally from above.

[0034] In another preferred embodiment, the device also has an illumination device for illuminating the container. Here, an incident light observation device is preferably provided, that is, the surface of the container is preferably observed from a direction that at least also has a component along the observation direction.

[0035] Preferably, at least one lighting device is synchronized with the image capture device. Thus, for example, a flash device can be provided. In a preferred embodiment, the lighting device has an LED light source.

[0036] In another advantageous embodiment, the sensor device is synchronized with the marking device and in particular triggered. This means that the marking device applies the marking to the container only for a very short time and during this time the marking is detected by the sensor device (or image recording device).

[0037] In another preferred embodiment, the image recording devices are also synchronized with one another. This means that the image recording devices particularly preferably record corresponding images of the container simultaneously. Small time deviations between the image recordings are also possible, wherein in this case, preferably, a movement of the container between the two recordings is additionally compensated and / or taken into account.

[0038] Preferably, the marking device is suitable and designed for contactless application of the marking to the container surface.

[0039] Temporary lighting can be understood in particular here to mean that the marking device can be switched on only at the time when the position determination is carried out, or can be switched on only at the time when the position determination is carried out, but it can also be illuminated during the period when the image recording device is not recording any spatially resolved images (or be switched off during the recording of images by the image recording device).

[0040] The marking recording and the image recording are preferably performed within a time interval of <10 ms. Particularly preferably, the object movement speed between the marking recording and the image recording is calculated in order to take the offset between the marking recording and the image recording into account in the position result.

[0041] Particularly preferably, the transport device performs a linear transport of the container. In a preferred embodiment, the transport device can have a conveyor belt and a conveyor chain on which the container is transported. However, other holding elements for the container are also conceivable, for example, holding elements that grip the container at the mouth of the container or under the support ring.

[0042] In another advantageous embodiment, the transport device transports the containers continuously. However, transport in a clocked manner is also conceivable.

[0043] In a preferred embodiment, at least four such image recording devices or cameras are provided, which record spatially resolved images of the container. However, a plurality of cameras, for example five or six cameras, may also be provided.

[0044] In another preferred embodiment, a plurality of image recording devices are provided, which are arranged on different planes one above the other, in particular relative to the longitudinal direction of the container, i.e., in particular also relative to the vertical direction. In this way, the container can be observed in different planes. Preferably, at least one position detection device as described above is also arranged in each of these planes.

[0045] The information that is particularly preferably applied on the surface and particularly preferably applied in the seam area is preferably applied by a pattern, a line, or a point or a plurality of points of projection. Here, the recording information can be filmed by the same camera that is also used to create a panoramic image.

[0046] This information can also be introduced into the same camera recording used to generate the panoramic image. The information can be evaluated and calculated with the help of appropriate processing steps. Or the information can be recorded in a second image recording with the same camera. If the container is moving continuously, the second image recording should preferably be carried out in a short time sequence.

[0047] However, it is also possible to record the information by means of an additional camera.

[0048] The number of additional cameras can also deviate from the number of cameras used to generate the panoramic image.

[0049] In another preferred embodiment, the marking device has a radiation device, which applies radiation to the surface. Here, the radiation device can be a laser, for example, but other light sources are also conceivable, such as LEDs, power LEDs, etc. It is also possible to use optical elements to apply the marking to the container surface, such as cylindrical lenses or the like.

[0050] In another preferred embodiment, the marking device is a pattern projector, which is suitable and designed for arranging a marking having at least one line or point on the surface of an object. However, it is also possible that the marking has a plurality of points or a plurality of lines. In this case, the lines can extend in a curved or straight manner. In addition, it is also possible that the lines extend parallel to the longitudinal direction of the object or container, but it is also conceivable that the lines extend at a specific angle, for example inclined or perpendicular to the longitudinal direction of the container. Advantageously, the light color of the pattern projector can be freely selected.

[0051] In another advantageous embodiment, the device has a processor device which is suitable and designed to combine the images recorded by the two image recording devices. The processor device is preferably intended to combine the corresponding images at a specific seam and particularly preferably to combine them without visible transitions. In this way, the subsequent analysis of the recorded panoramic image formed in this way can be simplified.

[0052] Preferably, the processor device also creates a panoramic image based on the at least two images taken.

[0053] Particularly preferably, the processor device is also controlled based on the data of the sensor device. This means that the panoramic image is also determined or created with the aid of the data recorded by the sensor device.

[0054] In a further advantageous embodiment, the position detection device comprises at least one second marking device which is suitable and designed to temporarily provide at least one surface section with a detectable marking. Here again, it can also be a laser or the like which applies such a marking to the surface of the container to be checked. As already mentioned, this is a temporary marking which is in particular also recorded by at least one image recording device.

[0055] Particularly preferably, the device also comprises more than two such marking devices, and particularly preferably, the number of marking devices corresponds to the number of image recording devices which record the container surface.

[0056] Here, these marking devices can be offset relative to each other in the circumferential direction of the container. Thus, for example, these marking devices can be offset from each other in the circumferential direction of the container by an angle between 70° and 110°.

[0057] In a further preferred embodiment, the position detection device comprises a second sensor device which is suitable and designed for detecting markings which are temporarily located on the surface.

[0058] Preferably, a respective sensor device is associated with each marking device.

[0059] In a further advantageous embodiment, at least one image recording device also functions simultaneously as a sensor device. Preferably, a plurality of image recording devices also function as sensor devices. It is conceivable that a control device is provided, which controls the image recording devices in such a way that first a recording is performed to detect the marking and then a second recording is performed for the image of the corresponding container to be recorded.

[0060] Thus, the information can also be processed with the same camera, in particular in a rapid sequence of photographs. In this case, preferably, the sequence has at least two recordings, wherein preferably at least one recording is recorded with temporarily additionally applied information or markings, and at least one recording does not contain additional information, such as, in particular, markings.

[0061] Therefore, the device preferably comprises at least one control device which controls the image recording device in such a way that a picture is first recorded in order to detect the marking and then a second picture is recorded in order to record the actual image of the container.

[0062] It would also be possible to use a suitable sensor device instead of a camera or in addition. The data of such a sensor can be provided in a data processing or data fusion manner to the unit that creates the panoramic strip.

[0063] In another advantageous embodiment, the position detection device comprises at least two sensor devices which are suitable and designed to detect the same marking. In this way, in particular, a junction or seam area between two images can be determined. Preferably, exactly two sensor devices detect the marking. In this way, the two sensor devices can control the corresponding image recording device or processor device, which then creates a panoramic image from the two recorded images.

[0064] The invention also relates to a method for inspecting an object, in particular a beverage container. Here, the object is transported along a predefined transport path by means of a transport device, and spatially resolved images of the object are captured from different directions by means of at least three image recording devices. Furthermore, a position detection device detects at least one position of at least one section of the object, and in particular at least one position of at least one surface section of the object.

[0065] According to the invention, the marking device temporarily marks a surface section of the object and the sensor device detects the marking.

[0066] Therefore, it is also proposed in the method that the applied marking is also detected by the sensor device. Particularly preferably, the marking is an illumination of the surface section of the container and in particular an illumination of a label arranged on the surface section of the container. Therefore, it is particularly preferred that the marking is temporarily applied to the label of the container to be checked. In particular, this is a label that has already been attached to the container. In a further preferred method, the marking is selected from a group of markings that includes lines, dots, grids, etc.

[0067] Particularly preferably, at least one image recording device detects the label.

[0068] In a further preferred method, the image recording device is simultaneously a sensor device which detects the marking.

[0069] In a preferred method, the containers are transported in a straight line.

[0070] In another preferred method, the containers are transported separately. Particularly preferably, a spacing is left between two consecutive containers, which spacing corresponds at least to the cross section of the container.

[0071] In another preferred embodiment, the processor device combines at least two captured images into an overall image and in particular a panoramic image. Particularly preferably, the processor device combines a plurality of captured images and in particular all images captured from a specific container into a panoramic image.

[0072] Preferably, the panoramic image consists of at least two captured images, wherein the panoramic image is preferably combined without forming visible seams.

[0073] In another preferred method, the reaction to the captured panoramic image determines whether the container is to be retained in the product flow or to be ejected from the product flow. Therefore, on the device side, the device can include an ejection device for ejecting individual containers.

[0074] Particularly preferably, the images captured by the image recording device are combined based on the sensor data.

[0075] In a further preferred method, two sensor devices are provided, which particularly preferably record images of two lighting devices. In this way, three image segments can be combined to form a panoramic image. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Further advantages and embodiments can be found in the drawings.

[0077] The accompanying drawings show:

[0078] Figure 1 is a schematic diagram of a device according to the present invention;

[0079] Figure 2 for the setting of the equipment according to the internal prior art of the applicant;

[0080] Figure 3 It is a view used to explain the problems in the prior art;

[0081] Figure 4 A diagram for explaining the problem on which the present invention is based;

[0082] Figure 5a to Figure 5c are three views for illustrating the captured images;

[0083] Figure 6a , Figure 6b , Figure 6c Three views for illustrating the method according to the present invention;

[0084] Figure 7 is another view of the method according to the present invention;

[0085] Figure 8a , Figure 8b are two further views for illustrating the method according to the invention; and

[0086] Fig. 9 FIG. 1 is another view of the method according to the present invention. DETAILED DESCRIPTION

[0087] Figure 1 A schematic diagram of a device 1 according to the invention is shown. The device 1 has a transport device 2 which transports objects to be inspected, such as in particular containers 10, along a transport path P. The transport path is straight here, but can also be curved, for example in the form of a circular arc, in another embodiment.

[0088] Reference numerals 12, 14, 16, 18 denote four image recording devices, which are arranged in the circumferential direction around the container 10 or the transport path of the container. Images of the container are recorded by means of the four image recording devices 12, 14, 16, 18, wherein the four image recording devices are in particular image cameras. These images are combined into an overall image by means of a processor device 20 (shown only schematically). This is in particular a panoramic image.

[0089] Figure 2 The structure of a device known from the applicant's internal prior art is shown. Six cameras 112, 114, 116, 118, 120 and 122 are provided here, which are arranged around the container 10 in the circumferential direction and in particular also observe the surface 10a of the container. In particular, a label can be placed on the surface 10a. In this way, the container 10 can be scanned all-round and different points can be checked for any existing label.

[0090] In this case, the image capturing devices or cameras 112 to 122 are arranged in one plane. More precisely, there is also at least one plane of the recognition unit. However, in addition, a plurality of image capturing devices can also be present in a plurality of planes arranged one above the other, i.e. in planes arranged one above the other in a direction perpendicular to the image plane. In this case, there can be an area B which can be captured by two adjacent image capturing devices.

[0091] It is known in the prior art to combine images recorded by image recording devices to form a panoramic image.

[0092] Within the scope of the following application, a method is now proposed for combining recorded individual images to form a panoramic image and in particular for seamlessly stitching the individual images.

[0093] However, if this information is not known precisely, you can try to find so-called feature points in the overlapping area of ​​the seam of adjacent images and align them.

[0094] Figure 3This shows the situation when the position of the container cannot be accurately determined or the shape of the container deviates from the assumed shape.

[0095] In both cases, the assumed spatial position of the container surface does not correspond to the actual position.

[0096] Figure 3 2 shows two adjacent image recording devices 12 and 14. Reference numeral O1 denotes a theoretically assumed surface of a container. Reference numeral O2 denotes the actual surface of a container.

[0097] The intersection point of the two rays S, indicated by reference numeral P3, does not lie on the actual surface (O2) of the container, but on the assumed surface O1. Reference numeral ds indicates the displacement of the actual position relative to the assumed position of the container. The intersecting angle line extended by a dashed line indicates the actual point of incidence on the real or actually existing container.

[0098] Figure 4 Another view is shown for illustrating the formation of a corresponding hypothetical panoramic image. Reference numeral E2 denotes a theoretically hypothetical surface of the container, and reference numerals E1 and E3 denote offset surfaces, respectively. In theory, the image contents of the two image recording devices would be seamlessly stitched together at the surface position of plane 2.

[0099] In the case of planes E1 and E3, two actual situations are drawn, in which the spatial position of the surface is not at the theoretical position. In the case of plane E3, the image content is reproduced twice at the intersection position; in the case of plane E1, the image content is "swallowed". How much is swallowed or how much is reproduced twice in the case of incorrect positioning depends on the angle of the corresponding drawing, the angle of incidence.

[0100] Figures 5a to 5c Three views are shown for reproducing an image. Figure 5a In the case shown, the surface position of the container corresponds to the theoretical position. Figure 5a In the case shown, the individual image sections A1 , A4 and A2 are reproduced realistically (and seamlessly), and there are also real or actually existing transitions in the cut-out regions.

[0101] exist Figure 5b In the case shown, a portion of the image, such as pixels A3 and A1, is reproduced doubly.

[0102] exist Figure 5c In the case shown, a portion of the image, such as Figure 5a and Figure 5b The image section A4 is shown being engulfed.

[0103] In all methods, the actual surface position in the seam region cannot be determined with sufficient accuracy, and as a result, duplicated texture points or missing texture points may occur in the seam region.

[0104] Containers for beverages have different precision. For example, beverage cans have a diameter of 0.2 mm and are at most elliptical in the same order of magnitude. In practice, deviations of 0.4 mm from the theoretical cylindrical surface occur.

[0105] Glass containers sometimes have diameter fluctuations of typically 1.5 to 2 mm. Here, too, the glass containers typically deviate from the cylindrical basic shape by 1.5 mm. Even with precisely determined positions, local deformations of the container surface are an important aspect that is neglected in the prior art.

[0106] It is also disadvantageous that the container position is measured via its contour in the camera image and adjacent containers must follow very large spacings in order to determine contours not covered by adjacent containers. With a predetermined output of xb / h, this would lead to correspondingly high conveyor belt speeds. However, this is undesirable in production facilities.

[0107] It is also disadvantageous that the container position is estimated at a distance via its diameter. Local deformations, diameter tolerances and measuring errors do not provide sufficiently accurate results. It is also disadvantageous that when determining the contour, the surface position is seemingly determined, which, however, can be a silhouette at best. In circularly symmetrical containers, total reflection occurs starting from a certain angle of incidence. Therefore, only the true contour can be estimated.

[0108] If the contour is to be determined by an image recording device or a camera, the position of the contour (line) does not coincide with the position of the seam line due to the arrangement of the camera.

[0109] Figure 6a The figure shows a first embodiment of the invention. Here again, two (only roughly indicated) image recording devices 12 and 14 are provided, which observe the container 10 from different directions. Reference numeral 22 denotes the above-mentioned marking device, which here applies a marking 30 to the container surface.

[0110] For example, this can be a vertical line 30 or the container is illuminated exactly with this line. Thus, the marking 30 can be applied at one point and preferably at multiple points and in such an orientation that the seam between the two images extends. Spatial reconstruction is not (mandatory) necessary. The laser line 30 determines the seam point in the image.

[0111] Figure 6b and 6cThe captured images are shown respectively. Thereby, the marking device is used to generate information, and the image 6c shows the corresponding information, in particular Figure 6c and Figure 6b The vertical line in .

[0112] More precisely, Figure 6b represents an image captured by the image capturing device 14, and Figure 6c : represents an image recorded by image recording device 12. Marking 30 appears in both images, so that the images can be joined at this point and in particular can be joined seamlessly in this case.

[0113] like Figure 7 As shown in , it is therefore proposed that the thread 30 is applied at the point / orientation at which the seam is subsequently stretched. Figure 7 The spatial configuration is performed as shown and the tangents in the adjacent images are redefined. The reference numeral S2 here denotes the actual intersection angle and the reference numeral S1 denotes the theoretical intersection angle of the images. The reference numerals again denote information or markings applied to the container, wherein the information is applied here at the actual container position.

[0114] Figure 8a Another design of the present invention is shown, and Figure 8b The correspondingly recorded camera image is shown. In this case, one or more lines are applied in the region of the seam, which lines extend in an orientation deviating from the seam. The lines can be recorded by two image recording devices (only one image recording device is shown here), wherein the spatial surface position can be determined by means of triangulation. After the spatial surface position has been determined, the intersection line in the camera image can be re-determined.

[0115] It is also possible to project one or more patterns onto the surface in order to reconstruct the surface position by means of strip projection. After the spatial surface position has been determined, the intersection line in the camera image can be determined again.

[0116] Fig. 9 Another embodiment of the present invention is shown. Here, a plurality of points are projected onto the surface in the vicinity of the joint (drawn by vertical dashed lines), and the surface position is reconstructed by means of strip projection. After the spatial surface position is determined, the intersection line in the camera image is re-determined.

[0117] By means of the marking device, for example a (straight and / or at least partially curved) line is visible in two adjacent camera images. Figure 6b and 6c, the area around the seam between image capture device 12 and image capture device 14 is shown for the two images shown in FIG. 1 . The projected seam line can be recognized in both images. This seam line can be determined in the images with the aid of simple image processing tools in order to, for example, stitch together the image portion of image capture device 12 located to the right of the information line and the image portion of image capture device 14 located to the left of the information line to form a seamless panorama.

[0118] This is possible without any additional image knowledge.

[0119] If a sequence of images is generated, of which at least one image contains temporal information and at least one image does not contain temporal information, a movement of the container on the transport path can be taken into account, which movement may have occurred during the recording. However, this movement may also be negligible.

[0120] Furthermore, additional information about the container, its shape or its position can often be obtained with the aid of the spatial reconstruction and the creation of the panoramic image can be improved.

[0121] For example, one or more points can be detected at the camera image and the coordinates can be determined at the camera image. Thus, for example, another camera image can be taken without projection to capture the real container surface or the real label.

[0122] In general, the markings shown here, such as seams, can run in a straight line, in a curved shape or in a zigzag shape.

[0123] Typically, after projecting the information in a suitable information form, different two downstream methods can be employed alternatively or together.

[0124] The seam can therefore be determined directly in the image information of two adjacent camera images using the projection information. It is also conceivable to determine the spatial surface position using the projection information and to determine an optimal seam line between two adjacent images based on the spatial surface position.

[0125] Seam lines can be applied before or after straightening the camera image.

[0126] A spatial reconstruction of the projected information onto the container is also performed with the help of the associated knowledge of the camera position (intrinsic and extrinsic parameters) in the viewing direction and the position and direction of the projection.

[0127] Furthermore, seam lines can also be used in the segmentation. Thus, for example, in areas that are not important or interesting for other tasks, the precise determination of seam lines may be of secondary importance and errors may be tolerated in such cases (e.g., in the case of a glass container with a label, the seam line is partially on the glass and partially on the label). However, the interesting or important position is the line that runs through the label area.

[0128] By means of the temporary application of the information described here, the surface position of the container can be determined precisely. It is thus possible for the first time to determine the intersection line between two adjacent images of the container individually for each container. In this way, it is prevented that image contents appear twice in the panoramic image or disappear in the panoramic image. It is thus possible with the invention to read the best before date completely and without error, for example, in the region of a seam.

[0129] The invention also eliminates the influence of the positioning of the container and / or the influence of production tolerances of the container.

[0130] The applicant reserves the right to claim the features disclosed in the present application documents as the essence of the present invention, as long as the features are novel with respect to the prior art, either individually or in combination. In addition, it should be noted that features that can be advantageous in themselves are also described in the various figures. It should be immediately recognized by those skilled in the art that a particular feature described in the figures can also be advantageous without the use of other features in the figures. In addition, those skilled in the art recognize that advantages can also be obtained by combining more features shown in individual or different figures.

[0131] Reference numerals list

[0132] 2 Transport device

[0133] 10 Container

[0134] 10a Container surface

[0135] 12 Image capture device

[0136] 14 Image capture device

[0137] 16 Image capture device

[0138] 18 Image capture device

[0139] 20 Processor Device

[0140] 22 Marking device

[0141] 30 mark, vertical line

[0142] 112 Camera

[0143] 114 Camera

[0144] 116 Camera

[0145] 118 Camera

[0146] 120 Camera

[0147] 122 Camera

[0148] A1 Image segment

[0149] A2 Image segment

[0150] A4 Image Section

[0151] E1 Staggered surfaces

[0152] Theoretical assumptions of E2 containers

[0153] E3 Staggered Surfaces

[0154] Theoretical assumed surface of the O1 container

[0155] O2 Surface of actual container

[0156] P Transportation Path

[0157] P3 Intersection point of rays

[0158] S-ray

[0159] Theoretical intersection angle of S1 image

[0160] Theoretical intersection angle of S2 image

[0161] ds Displacement of actual position

[0162] Area B.

Claims

1. A device (1) for inspecting a beverage container, the device comprising: A transport device (2) transporting an object (10) along a predetermined transport path; at least three image recording devices (12, 14, 16, 18) positioned so that they record spatially resolved images of the transported object (10) from different directions, wherein the device (1) comprises a position detection device (20) for detecting at least one position of at least one section of the object (10), It is characterized in that The position detection device (20) comprises: a first marking device (22), the first marking device is suitable for and determined to: temporarily set a mark (30) on at least one surface segment of the object (10); and a first sensor device, the first sensor device is suitable for and determined to: detect the mark (30) temporarily located on the surface (10a), the device (1) comprises a processor device (8), the processor device is suitable for and determined to combine images captured by two image capture devices at a seam, wherein the seam is obtained by detecting the mark by the first sensor device, thereby determining the seam area between the two images, and wherein the processor device creates a panoramic image based on at least two of the captured images, and the image capture device detects the tag.

2. The device (1) according to claim 1, It is characterized in that The first marking device (22) comprises a radiation device (32) for applying radiation to the surface (10a).

3. The device (1) according to claim 1, It is characterized in that The first marking device (22) is suitable and designed for arranging a marking (30) having at least one line on the surface (10a) of the object (10).

4. The device (1) according to claim 1, It is characterized in that The position detection device (20) comprises at least one second marking device (24) which is suitable and designed for temporarily providing at least one surface section with a detectable marking (30).

5. The device (1) according to claim 1, It is characterized in that The position detection device (20) comprises a second sensor device which is suitable and designed for detecting the marking temporarily located on the surface (10a).

6. The device (1) according to claim 1, It is characterized in that At least one image recording device (12, 14, 16, 18) also functions as a sensor device.

7. The device (1) according to any one of claims 1 to 6, It is characterized in that The position detection device (20) comprises at least two sensor devices which are suitable and designed to detect the same marking (30).

8. A method for inspecting beverage containers, wherein an object (10) is transported along a predefined transport path by means of a transport device (2), wherein spatially resolved images of the object (10) are recorded from different directions by means of at least three image recording devices (12, 14, 16, 18), and wherein a position detection device (20) detects at least one position of at least one section of the object (10), It is characterized in that A marking device (22) temporarily marks a surface section of the object (10) and a sensor device detects the mark, and a processor device (8) is provided, which is suitable for and determined to combine images captured by two image capture devices at a seam, wherein the seam is obtained by the sensor device detecting the mark, thereby determining a seam area between the two images, and wherein the processor device creates a panoramic image based on at least two of the captured images, and the image capture device detects the tag.

9. The method according to claim 8, It is characterized in that The mark is selected from the group consisting of a line, a dot or a grid.

Images