Method and device for producing an image of a bottom of a glass vessel

Abstract

A method and a device for producing an image of a bottom of a glass vessel having an axis use a matrix camera having pixels arranged in a plurality of rows and a plurality of columns. A series of individual photographs of strip-shaped regions of the bottom is captured by means of the matrix camera; mutually adjacent individual photographs partly overlap. A light source is used to transmit light through the bottom of the glass vessel during an individual photograph. Between different individual photographs, the glass vessel is rotated about its axis relative to the matrix camera. A digital image of the bottom of the glass vessel is composed from the series of individual photographs, the individual photographs being arranged at an angle relative to each other.

Description

[0001] Method and device for generating an image of a bottom of a glass vessel

[0002] The method and device relate to the inspection of glass vessels that have a bottom and an axis. However, the glass vessels do not need to be rotationally symmetrical.

[0003] Particularly during the industrial production of glass containers, defects can occur, albeit to a minor extent, in the base of the containers, such as defects, cracks, discoloration, foreign material, or air inclusions. If such defects are detected in a timely manner, the defective containers can be removed from the manufacturing process.

[0004] EP 2 434276 B1, for example, discloses an inspection method for checking transparent or translucent containers for defects such as cracks, fissures, bubbles, or the like. The containers are continuously conveyed along a conveying direction by a conveyor system. Each container passes through an inspection station, where a non-contact inspection of at least one selected container area of ​​each container takes place.

[0005] In practice, difficulties arise particularly when the bottom of a vessel is to be inspected while the vessel is standing with its bottom on a supporting element, such as a shelf.

[0006] The object of the invention is to show a way in which a bottom of a glass vessel can be inspected in an improved manner.

[0007] This object is achieved independently by a method according to claim 1 or by a device according to claim 10. Advantageous further developments are the subject of the dependent claims.

[0008] In a first aspect of the invention, a method is provided for generating an image of the bottom of a glass vessel having an axis. The axis of the glass vessel is typically substantially perpendicular to the bottom of the glass vessel, but the bottom may have a curvature, and the glass vessel need not be rotationally symmetrical about the axis. In the context of the invention, the term "glass vessel" includes not only vessels that actually comprise glass as their material, but also vessels that comprise transparent or at least translucent plastics as their material.

[0009] The method uses a matrix camera with pixels arranged in a plurality of rows and columns. Unlike previously used line-scan cameras, this camera has a flat recording area. This can be a CCD camera or a CMOS camera, for example, a so-called high-speed camera.

[0010] Using this matrix camera, a series of individual images is captured, each of which captures striped areas of the bottom of the glass vessel. These striped areas can be rectangular, but can also have other shapes, such as square or elliptical. Adjacent individual images overlap in sections.

[0011] During a single exposure, the bottom of the glass vessel is illuminated by a light source located on the opposite side of the bottom of the glass vessel from the matrix camera. The light source can be a pulsed light source, preferably operated synchronously with the acquisition of the individual exposures by the matrix camera.

[0012] To allow the light from the light source to pass through the container base into the optical input of the camera, a recess in the aforementioned support element is advantageous. The light source is located below it, and the container stands above it. In this case, the length of the recess should be at least large enough to extend from the center of the container base to beyond the edge of the container. However, it has proven advantageous if the length of the recess is greater than the entire diameter of the container, so that the recorded illuminated image strip spans the entire base area. The width of the recess should be selected so that the container cannot fall through and still stands securely on at least one, but preferably on both sides of the recess.

[0013] Between individual images, the glass vessel is rotated around its axis relative to the matrix camera. Finally, a digital image of the bottom of the glass vessel is assembled from the series of individual images, arranged at an angle relative to each other.

[0014] During rotation, the rotation axis may also shift relative to the container axis, so that the individual images not only exhibit the desired angle of rotation relative to the previous image, but the container may also be shifted in the X and Y directions within the image. When compositing the individual images, it may therefore be necessary to consider a translational change in addition to the rotational change.

[0015] The rotation and translation of the containers from one image to the next should be as large as possible to ensure that the image area no longer overlaps. It has proven advantageous to have a certain overlap, for example, 10% or up to 50% of the image area, in both images. This overlap allows the individual images to be perfectly aligned based on features that are simultaneously visible in the adjacent images.

[0016] The digital image can be stitched together during the creation of the series of individual images or after the entire series of individual images has been created. A series of individual images consists of at least two individual images, but preferably of at least three, at least four, at least five, at least 10, or even at least 15, 20, 30, or 50 individual images.

[0017] This method offers considerable advantages over conventional methods in which the bottom of glass vessels was scanned using a line scan camera - i.e. a camera with only a single image line, the images of which were stitched together in the form of a developed image without any relative rotation. Defects in the bottom of the glass vessel were sometimes extremely distorted or not displayed at all using the conventional method. The method according to the invention, on the other hand, allows a comparatively rapid generation of an image that is largely free of distortion and, above all, completely captures the bottom of the glass vessel. This image can then be read out for defects in the glass vessel either visually by a user and / or automatically using suitable image processing software.

[0018] The angular arrangement of the individual images relative to one another is preferably carried out as precisely as possible in the same way as the areas of the ground captured by the individual images are arranged at an angle to one another.

[0019] It is useful to use software to identify special spots in the individual images and to superimpose the special spots onto one another to create a digital image of the bottom of the glass vessel. Such special spots can be defects, such as the aforementioned defects, cracks, discoloration, or air inclusions (bubbles). Special spots can also be structures deliberately incorporated into the bottom of the glass vessel, such as text, grooves, or markings. The software, e.g., including an image capture module, can be configured to identify such special spots and to suitably join the individual images, ensuring the best possible match between the special spots.

[0020] The stitching of the digital image of the bottom of the glass vessel may involve rotation, linear translation, and / or stretching or compressing one or more individual images. These measures may, for example, be aimed at creating the best possible superimposition of identified special spots. Artificial intelligence (AI) can be used to stitch the digital image of the bottom of the glass vessel, enabling optimization of the stitching of the digital image through suitable self-learning processes.

[0021] It is advantageous if the digital image of the bottom of the glass vessel is stitched taking into account the relative rotation between the glass vessel and the matrix camera between each two individual images. The magnitude of this relative rotation between the glass vessel and the matrix camera between each two individual images can be known, constant, and / or predetermined by the rotational movement. Using the magnitude of the predetermined or performed relative rotation as an input variable in the software used to stitch the digital image reduces the computing power and time required to stitch the digital image.

[0022] The actual relative rotation between the glass vessel and the matrix camera between each two individual images can be, for example, 1 to 15°, preferably 2 to 12°. Amounts outside these ranges are also conceivable. The following applies: The greater the angle of relative rotation between two individual images, the fewer the number of individual images required for an image (and correspondingly lower the computing power), but also the lower the achievable resolution.

[0023] Preferably, the base of the glass vessel rests on a support structure, in particular a translucent surface, optionally with at least one recess, while the individual images are being taken. For example, the light source can be located below the support structure or surface, while the camera looks down onto the base of the glass vessel. If the base of the glass vessel rests on a support structure, this has the advantage that the base of the glass vessel is always in the same plane while the series of individual images is being taken. This makes it easier to focus the individual images and thus improves the resolution of the digital image of the base. However, another variant is also conceivable, in which a (particularly axially symmetrical) glass vessel with a horizontal axis is mounted on two rotating, driven rollers, which are also horizontal, as a support structure.

[0024] It has proven advantageous if the area of ​​the bottom of the glass vessel captured by a single image is longer than the diameter of the glass vessel. This allows the entire bottom of the glass vessel to be fully captured after a rotation of the glass vessel of less than 180°.

[0025] An alternative design could be a pickup extending from the center of the container to the diameter in order to capture the entire bottom during a rotation of less than 360°.

[0026] It may be expedient if the area of ​​the bottom of the glass vessel captured by an individual image has a width of 10% to 30% of the diameter of the glass vessel, preferably a width of 15% to 25% of the diameter of the glass vessel. Without the file size of individual images becoming excessively large, with a minimum width of 10% or 15% of the diameter, the probability is sufficiently high that sufficient features are located in the overlapping area of ​​adjacent individual images to facilitate the compositing of the individual images. In a second aspect, the invention relates to a device for generating an image of a bottom of a glass vessel having an axis.The device comprises a matrix camera whose pixels are arranged in a plurality of rows and a plurality of columns, as well as a support structure for supporting the glass vessel and a light source for illuminating the bottom of the glass vessel, wherein the light source is arranged on the side of the support structure opposite the matrix camera. The device further comprises a drive for generating a rotation of the glass vessel held by the support structure relative to the matrix camera and a memory for storing a series of individual images of strip-shaped regions of the bottom of the glass vessel taken by the matrix camera. Furthermore, the device comprises an evaluation unit configured to assemble a digital image of the bottom of the glass vessel from the series of individual images, with the individual images arranged at an angle relative to one another.These measures result in the advantages described in the introduction with regard to the first aspect.

[0027] The evaluation unit is preferably configured to detect special spots in the individual images and to combine the individual images by superimposing the special spots. These special spots can, as mentioned above, be defects (e.g., cracks or air pockets) or deliberately introduced special spots such as text, grooves, or markings. The individual images are preferably combined in such a way that the special spots are optimally superimposed in shape, size, and orientation. To achieve this superimposition, the evaluation unit can comprise software that incorporates artificial intelligence (AI) and implements self-learning processes.

[0028] It is useful if the evaluation unit is configured to perform rotation, linear translation, and / or stretching or compressing of an individual image to compose the digital image of the bottom of the glass vessel. Each or several of these measures serve to optimize the superimposition of the individual images and the special spots they contain.

[0029] The light source used to illuminate the bottom of the glass vessel can be operated in a pulsed mode to enable particularly high light output during individual exposures while reducing energy consumption. Preferably, the light source can be operated in a synchronized manner with the matrix camera, so that the pulsing of the light source is synchronized with the acquisition of the individual exposures.

[0030] It has proven advantageous if the support structure is designed to perform a relative rotation between the glass vessel and the matrix camera by an angle of 1 to 15° between every two individual images, preferably by an angle of 2 to 12°. The smaller the amount of relative rotation, the higher the resolution achievable with the digital image. The support structure could, for example, be a translucent surface on which the base of the glass vessel rests while the individual images are taken. Alternatively, the support structure can have a group of horizontally mounted rollers on which the glass vessel is mounted horizontally and rotated by the drive of one of the rollers.

[0031] In a third aspect, the invention relates to a computer program product which, when run on a computer, is configured to compose a digital image of the bottom of a glass vessel from a series of individual images of the bottom of a glass vessel, with the individual images arranged at an angle relative to one another. The computer can be part of the evaluation unit of the device. The computer program product can comprise an image evaluation module configured to detect special locations on the bottom in the individual images, and / or the computer program product can comprise artificial intelligence (AI) configured to optimize the superimposition of the individual images. Preferably, the computer program product is configured to optimize the superimposition of the individual images for compiling the digital image with regard to the special locations detected in the individual images.

[0032] Elements or properties described in connection with one of the aspects (method, device or computer program product) can also be implemented individually or in combination in one of the other two aspects in the context of the invention.

[0033] In the following, the invention is further explained using an embodiment with reference to the figures.

[0034] Figure 1 shows a schematic plan view of a device for inspecting vessels.

[0035] Figure 2 shows a schematic sectional view of a device for generating an image of the vessel bottom according to an embodiment, the section being indicated by 11 in Figure 1.

[0036] Figure 3 shows a schematic representation of several individual images of the bottom of the glass vessel.

[0037] Figure 4 shows a schematic representation of a composite image of the bottom of the glass vessel.

[0038] Figure 1 shows a schematic plan view of a device 1 for inspecting containers 3. As shown in Figure 2, the containers 3 are, for example, glass bottles with a bottom 5 and a side wall 7. Alternatively, the containers 3 can be, for example, other types of packaging glass, e.g. jam or preserving jars.

[0039] As shown in Figure 1, the device 1 comprises a transport device 9 for transporting the containers 3 along a transport direction 11. In the illustrated embodiment, the transport device 9 has a star wheel 13 which transports the containers 3 along a circular path. The star wheel 13 comprises holding elements 15 which are arranged one behind the other along a circumferential direction of the star wheel 13. The containers 3 are transferred from a transfer station 17 to the star wheel 13 by being placed between adjacent holding elements 15 of the star wheel 13. By rotating the star wheel 13, the containers 3 are conveyed along the transport direction 11. During conveyance, the containers 3 are pushed by the holding elements 15 of the star wheel 13 over a transport surface 19 of the transport device 9. The transport of the containers 3 along the transport direction 11 takes place in a clocked manner.After the vessels 3 have been inspected in the device 1, they are removed from the transport device 9 by a removal station 21 located downstream of the transfer station 17 with respect to the transport direction 11.

[0040] With respect to the transport direction 11 between the transfer station 17 and the removal station 21, an inspection station 23 is provided, at which the bottom 5 of the vessel 3 present in the inspection station 23 is examined for imperfections or defects. During the inspection of a vessel 3 by the inspection station 23, the star wheel 13 is preferably stationary. Therefore, preferably, no transport of the vessel 3 along the transport direction 11 takes place during this time.

[0041] During the inspection of a vessel 3 in the inspection station 23, the vessel 3 is in an inspection position. In the inspection position, the vessel 3 is in contact with a drive or rotating device 25. In the inspection position, the vessel 3 is rotated by the rotating device 25 about an axis 27 of the vessel (see Figure 2) along a rotation direction 29.

[0042] Figure 2 shows a sectional view in the area of ​​the inspection station 23 along the section indicated by 11 in Figure 1. A device 24 according to the invention for generating an image of the bottom 5 of the glass vessel 3 is arranged at the inspection station 23. The device 24 and the most important components of this device are shown in Figure 2.

[0043] The vessel 3 shown in Figure 2 is in the inspection position. In the inspection position, the vessel 3 rests with its base 5 on a support structure 30. In the illustrated embodiment, the support structure 30 is inserted into a receptacle of the transport surface 19. According to embodiments, the support structure 30 can be inserted interchangeably into the transport surface 19. Alternatively, the support structure 30 can be formed integrally with the transport surface 19. The transport surface 19 and the support structure 30 can have flush upper surfaces so that the bottle 3 can be pushed from the transport surface 19 onto the support device 30 by the star wheel 13. Alternatively to the illustrated embodiment, the rotating device 25 can be a drive designed to generate a rotation of the support structure 30 about the axis 27 of the vessel 3.

[0044] A matrix camera 39 is arranged above the support structure with a vertically downward viewing direction. The vessel 7 is centered with its axis 27 essentially relative to the viewing direction of the matrix camera 39, which is directed from above through the opening of the vessel 7 onto its bottom 5. The matrix camera 39 is characterized in that its image points (pixels) 40, as shown in Figure 3, are arranged in a plurality of rows Z and a plurality of columns S, i.e., on a surface (instead of just in a single row).

[0045] On the side of the support structure 30 opposite the matrix camera 39, i.e., below the support structure 30 in the illustrated embodiment, a light source 37 is arranged. The light source 37 serves to illuminate the bottom 5 of the glass vessel. For this purpose, the support structure 30 can, for example, have a translucent surface 31 so that the light emitted by the light source 37 can penetrate the bottom 5 of the vessel 3. One or more recesses 31a through which light can pass can be present in the support structure 30 or the translucent surface 31. The light source 37 can be a pulsed light source, e.g., a stroboscopic light source. In this case, the emission of its light pulses can be synchronized with the operation of the matrix camera 39, e.g., by a control (not shown) of the device 24.

[0046] On the camera side, one variant can utilize an optics system with an integrated beam splitter and two attached cameras 39. One of the cameras 39 is arranged axially, as shown in Figure 2, while the other is mounted laterally at an angle of 90°. The light source 37 is equipped with a linear polarizing filter, and the camera optics with a linearly polarizing beam splitter. One camera 39 thus sees a bright image, while the other camera 39 normally sees nothing because the polarizing filters are arranged crossed. However, if there is a stress-bearing inclusion (defect) in the bottle base 5, the polarization plane is rotated, and the second camera 39 sees the stress source as a bright spot. The two cameras 39 thus serve for normal ground control and stress control. The use of a station 23 with only one camera without polarization evaluation is also conceivable as an alternative. The use of image sensors with a front-mounted polarizing filter is also possible.

[0047] Figure 3 shows a schematic representation of several individual images E taken by the matrix camera 39. Due to the orientation of the matrix camera 39 and the arrangement of its pixels 40 in several rows Z and columns S, each individual image E consists of a recording of a strip-shaped area B of the base 5 of the glass vessel 3. In Figure 3, the recorded area B of the base 5 is the intersection between the circular base 5 of the glass vessel 3 and the total area of ​​the individual image E. Each individual image E covers a specific length L and a specific width b. The length L of the area B of the base 5 captured by an individual image E is greater than a diameter I of the glass vessel 3, while the area B of the base 5 of the glass vessel 3 captured by an individual image E has a width b of approximately 10% to 30% of the diameter I of the glass vessel 3.While a series of individual images E is being taken of the bottom 5 of a glass vessel 3, a relative rotation of the vessel 3 about its axis 27 occurs between different individual images E. The relative rotation between two individual images E can occur by an angle of, for example, 1° to 15°, preferably by an angle of 2° to 12°. The rotation by the angle α is generated by the rotating device 25.

[0048] The device 24 comprises an evaluation unit 41, which can be integrated into the matrix camera 39 or connected to the matrix camera 39. The evaluation unit 41 comprises a memory 42 for storing a series of individual images E and a computer 43 on which a computer program product 44 is installed. The evaluation unit 41, or more specifically the computer program product 44 installed on it, is configured to compile a digital image of the bottom 5 from a series of individual images E of a bottom 5 of the glass vessel 3. Figure 3 indicates how this can be done:

[0049] In the base 5 of the glass vessel 3 there are a plurality of special spots 45. The special spots 45 can be deliberately introduced into the base 5, e.g., circumferential indentations 45a, or an undesired defect 45b, e.g., a bubble or a crack. An image recognition module of the computer program product 44 is configured to detect such defects 45 in the individual images E. The evaluation device 41 is then configured to manipulate the individual images E in such a way that an optimal superposition of the special spots 45 in the respective individual images E is achieved. The manipulation can comprise rotating the respective individual images E (e.g., but not necessarily, about the axis 27 of the vessel 3), translating the individual images E in their longitudinal and / or transverse direction, and / or stretching or compressing the respective individual images E.

[0050] When all individual images E of a series have been processed by the evaluation unit 41, it has generated a digital image A of the bottom 5 of the glass vessel 3, as shown in Figure 4. The digital image A is composed of the respective individual images E, with the individual images E arranged at an angle relative to one another. As a result, this does not produce a "development" of the bottom 5 with corresponding distortions, but rather a distortion-free image of the bottom 5 of the glass vessel 3.

[0051] To facilitate the evaluation and compositing of the image A, the evaluation unit 41 can consider as an input variable the angle a by which the glass vessel 3 is rotated relative to the matrix camera 39 between two individual images E. This input variable makes it easier for the evaluation unit 41 to compose the digital image A, since the probability of the need to rotate the individual images E is reduced.

[0052] If the device 1, the inspection station 23, or the device 24 has a display 46 (see Figure 2), the digital image A can be displayed there. Alternatively, the digital image A can be evaluated mechanically. If defects 45b are detected, the corresponding glass vessel 3 can be manually or automatically ejected.

[0053] Based on the illustrated embodiments and the appended claims, the invention can be modified in various ways. One possibility, for example, is to capture and inspect individual images (visually or mechanically) before, or even without, composing a digital image (A) of the bottom (5) of the glass vessel (3) from multiple images.

Claims

Claims 1. A method for generating an image (A) of a bottom (5) of a glass vessel (3) having an axis (27), wherein a matrix camera (39) having pixels (40) arranged in a plurality of rows (Z) and a plurality of columns (S) is provided, wherein a series of individual images (E) of strip-shaped regions (B) of the bottom (5) of the glass vessel (3) is taken by means of the matrix camera (39), wherein adjacent individual images (E) overlap in sections, wherein during an individual image (E), the bottom (5) of the glass vessel (3) is illuminated by means of a light source (37) arranged on the side of the bottom (5) of the glass vessel (3) opposite the matrix camera (39), wherein between different individual images (E), a rotation of the glass vessel (3) about its axis (27) relative to the matrix camera (39) takes place,and wherein a digital image (A) of the bottom (5) of the glass vessel (3) is composed from the series of individual images (E) with an angular arrangement of the individual images (E) relative to one another.

2. Method according to claim 1, wherein, for assembling the digital image (A) of the bottom (5) of the glass vessel (3), software (44) is configured to recognize special points (45) in the individual images (E) and to join the individual images (E) together by superimposing the special points (45).

3. Method according to one of the preceding claims, wherein a rotation, linear displacement and / or stretching or compressing of an individual image (E) takes place to assemble the digital image (A) of the bottom (5) of the glass vessel (3).

4. Method according to one of the preceding claims, wherein the composing of the digital image (A) of the bottom (5) of the glass vessel (3) takes into account the relative rotation between the glass vessel and the matrix camera (39) between each two individual images (E).

5. Method according to one of the preceding claims, wherein between each two individual images (E) a relative rotation between the glass vessel (3) and the matrix camera (39) takes place by an angle (α) of 1 to 15°, preferably by an angle (α) of 2 to 12°.

6. Method according to one of the preceding claims, wherein the bottom (5) of the glass vessel (3) stands on a support structure (30), in particular a light-permeable support surface (31), during the recording of the individual images (E).

7. Method according to one of the preceding claims, wherein the region (B) of the bottom (5) of the glass vessel (3) captured by an individual image (E) has a greater length (L) than a diameter (I) of the glass vessel (3).

8. Method according to one of the preceding claims, wherein the region (B) of the bottom (5) of the glass vessel (3) captured by the individual images (E) is penetrated by the axis (27) of the glass vessel (3).

9. Method according to one of the preceding claims, wherein the region (B) of the bottom (5) of the glass vessel (3) captured by an individual image (E) has a width of 10% to 30% of the diameter (I) of the glass vessel (3), preferably a width of 15% to 25% of the diameter (I) of the glass vessel (3).

10. A device (24) for generating an image (A) of a bottom (5) of a glass vessel (3) having an axis (27), comprising a matrix camera (39) with pixels (40) arranged in a plurality of rows (Z) and a plurality of columns (S), a support structure (30) for supporting the glass vessel (3), a light source (37) for illuminating the bottom (5) of the glass vessel (3), wherein the light source (37) is arranged on the side of the support structure (30) opposite the matrix camera (39), wherein a drive (25) is provided for generating a rotation of the glass vessel (3) held by the support structure (30) relative to the matrix camera (39), wherein a memory (43) is provided for storing a series of individual images (E) of strip-shaped regions (B) of the bottom (5) of the glass vessel (3) taken by the matrix camera (39), and wherein the device (24) has an evaluation unit (41) which is designed to assemble a digital image (A) of the bottom (5) of the glass vessel (3) from the series of individual images (E) with an angular arrangement of the individual images (E) relative to one another.

11. Device according to claim 10, wherein the evaluation unit (41) is configured to detect special points (45) in the individual images (E) and to join the individual images (E) together by superimposing the special points (45).

12. Device according to one of claims 10 or 11, wherein the evaluation unit (41) is configured to carry out a rotation, linear displacement and / or a stretching or compression of an individual image in order to assemble the digital image (A) of the bottom (5) of the glass vessel (3).

13. Device according to one of claims 10 to 12, wherein the light source (37) can be operated in a pulsed manner, wherein preferably the light source (37) can be operated in a synchronized manner with the recording of the individual images (E) by the matrix camera (39).

14. Device according to one of claims 10 to 13, wherein the support structure (30) is adapted to carry out a relative rotation between the glass vessel (3) and the matrix camera by an angle (α) of 1 to 15°, preferably by an angle (α) of 2 to 12°, between each two individual recordings (E).

15. Computer program product (44) which, when run on a computer (43), is designed to assemble a digital image (A) of the bottom (5) of the glass vessel (3) from a series of individual images (E) of a bottom (5) of a glass vessel (3) with an angular arrangement of the individual images (E) relative to one another.

Images